9 .Nd User interface for firewall, traffic shaper, packet scheduler,
12 .Ss FIREWALL CONFIGURATION
21 .Op Ar rule | first-last ...
29 .Brq Cm delete | zero | resetlog
33 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
37 .Ar number Cm to Ar number
39 .Cm set swap Ar number number
45 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
48 .Brq Cm firewall | altq | one_pass | debug | verbose | dyn_keepalive
51 .Oo Cm set Ar N Oc Cm table Ar name Cm create Ar create-options
53 .Oo Cm set Ar N Oc Cm table
57 .Oo Cm set Ar N Oc Cm table Ar name Cm modify Ar modify-options
59 .Oo Cm set Ar N Oc Cm table Ar name Cm swap Ar name
61 .Oo Cm set Ar N Oc Cm table Ar name Cm add Ar table-key Op Ar value
63 .Oo Cm set Ar N Oc Cm table Ar name Cm add Op Ar table-key Ar value ...
65 .Oo Cm set Ar N Oc Cm table Ar name Cm atomic add Op Ar table-key Ar value ...
67 .Oo Cm set Ar N Oc Cm table Ar name Cm delete Op Ar table-key ...
69 .Oo Cm set Ar N Oc Cm table Ar name Cm lookup Ar addr
71 .Oo Cm set Ar N Oc Cm table Ar name Cm lock
73 .Oo Cm set Ar N Oc Cm table Ar name Cm unlock
75 .Oo Cm set Ar N Oc Cm table
79 .Oo Cm set Ar N Oc Cm table
83 .Oo Cm set Ar N Oc Cm table
87 .Oo Cm set Ar N Oc Cm table
90 .Ss DUMMYNET CONFIGURATION (TRAFFIC SHAPER AND PACKET SCHEDULER)
92 .Brq Cm pipe | queue | sched
98 .Brq Cm pipe | queue | sched
99 .Brq Cm delete | list | show
108 .Ss STATEFUL IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
110 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm create Ar create-options
112 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm config Ar config-options
114 .Oo Cm set Ar N Oc Cm nat64lsn
119 .Oo Cm set Ar N Oc Cm nat64lsn
123 .Oo Cm set Ar N Oc Cm nat64lsn Ar name Cm stats Op Cm reset
124 .Ss STATELESS IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
126 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm create Ar create-options
128 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm config Ar config-options
130 .Oo Cm set Ar N Oc Cm nat64stl
134 .Oo Cm set Ar N Oc Cm nat64stl
138 .Oo Cm set Ar N Oc Cm nat64stl Ar name Cm stats Op Cm reset
139 .Ss XLAT464 CLAT IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
141 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm create Ar create-options
143 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm config Ar config-options
145 .Oo Cm set Ar N Oc Cm nat64clat
149 .Oo Cm set Ar N Oc Cm nat64clat
153 .Oo Cm set Ar N Oc Cm nat64clat Ar name Cm stats Op Cm reset
154 .Ss IPv6-to-IPv6 NETWORK PREFIX TRANSLATION
156 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm create Ar create-options
158 .Oo Cm set Ar N Oc Cm nptv6
162 .Oo Cm set Ar N Oc Cm nptv6
166 .Oo Cm set Ar N Oc Cm nptv6 Ar name Cm stats Op Cm reset
167 .Ss INTERNAL DIAGNOSTICS
174 .Ss LIST OF RULES AND PREPROCESSING
187 utility is the user interface for controlling the
191 traffic shaper/packet scheduler, and the
192 in-kernel NAT services.
194 A firewall configuration, or
198 numbered from 1 to 65535.
199 Packets are passed to the firewall
200 from a number of different places in the protocol stack
201 (depending on the source and destination of the packet,
202 it is possible for the firewall to be
203 invoked multiple times on the same packet).
204 The packet passed to the firewall is compared
205 against each of the rules in the
208 (multiple rules with the same number are permitted, in which case
209 they are processed in order of insertion).
210 When a match is found, the action corresponding to the
211 matching rule is performed.
213 Depending on the action and certain system settings, packets
214 can be reinjected into the firewall at some rule after the
215 matching one for further processing.
217 A ruleset always includes a
219 rule (numbered 65535) which cannot be modified or deleted,
220 and matches all packets.
221 The action associated with the
227 depending on how the kernel is configured.
229 If the ruleset includes one or more rules with the
236 the firewall will have a
238 behaviour, i.e., upon a match it will create
240 i.e., rules that match packets with the same 5-tuple
241 (protocol, source and destination addresses and ports)
242 as the packet which caused their creation.
243 Dynamic rules, which have a limited lifetime, are checked
244 at the first occurrence of a
249 rule, and are typically used to open the firewall on-demand to
250 legitimate traffic only.
257 for all packets (not only these matched by the rule) but
264 .Sx STATEFUL FIREWALL
267 Sections below for more information on the stateful behaviour of
270 All rules (including dynamic ones) have a few associated counters:
271 a packet count, a byte count, a log count and a timestamp
272 indicating the time of the last match.
273 Counters can be displayed or reset with
277 Each rule belongs to one of 32 different
281 commands to atomically manipulate sets, such as enable,
282 disable, swap sets, move all rules in a set to another
283 one, delete all rules in a set.
284 These can be useful to
285 install temporary configurations, or to test them.
288 for more information on
291 Rules can be added with the
293 command; deleted individually or in groups with the
295 command, and globally (except those in set 31) with the
297 command; displayed, optionally with the content of the
303 Finally, counters can be reset with the
310 The following general options are available when invoking
312 .Bl -tag -width indent
314 Show counter values when listing rules.
317 command implies this option.
319 Only show the action and the comment, not the body of a rule.
323 When entering or showing rules, print them in compact form,
324 i.e., omitting the "ip from any to any" string
325 when this does not carry any additional information.
327 When listing, show dynamic rules in addition to static ones.
329 When listing, show only dynamic states.
330 When deleting, delete only dynamic states.
332 Run without prompting for confirmation for commands that can cause problems if misused,
335 If there is no tty associated with the process, this is implied.
338 command with this flag ignores possible errors,
339 i.e., nonexistent rule number.
340 And for batched commands execution continues with the next command.
342 When listing a table (see the
344 section below for more information on lookup tables), format values
346 By default, values are shown as integers.
348 Only check syntax of the command strings, without actually passing
351 Try to resolve addresses and service names in output.
353 Be quiet when executing the
363 This is useful when updating rulesets by executing multiple
367 .Ql sh\ /etc/rc.firewall ) ,
368 or by processing a file with many
370 rules across a remote login session.
371 It also stops a table add or delete
372 from failing if the entry already exists or is not present.
374 The reason why this option may be important is that
375 for some of these actions,
377 may print a message; if the action results in blocking the
378 traffic to the remote client,
379 the remote login session will be closed
380 and the rest of the ruleset will not be processed.
381 Access to the console would then be required to recover.
383 When listing rules, show the
385 each rule belongs to.
386 If this flag is not specified, disabled rules will not be
389 When listing pipes, sort according to one of the four
390 counters (total or current packets or bytes).
392 When listing, show last match timestamp converted with ctime().
394 When listing, show last match timestamp as seconds from the epoch.
395 This form can be more convenient for postprocessing by scripts.
397 .Ss LIST OF RULES AND PREPROCESSING
398 To ease configuration, rules can be put into a file which is
401 as shown in the last synopsis line.
405 The file will be read line by line and applied as arguments to the
409 Optionally, a preprocessor can be specified using
413 is to be piped through.
414 Useful preprocessors include
420 does not start with a slash
422 as its first character, the usual
424 name search is performed.
425 Care should be taken with this in environments where not all
426 file systems are mounted (yet) by the time
428 is being run (e.g.\& when they are mounted over NFS).
431 has been specified, any additional arguments are passed on to the preprocessor
433 This allows for flexible configuration files (like conditionalizing
434 them on the local hostname) and the use of macros to centralize
435 frequently required arguments like IP addresses.
436 .Ss TRAFFIC SHAPER CONFIGURATION
442 commands are used to configure the traffic shaper and packet scheduler.
444 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
445 Section below for details.
447 If the world and the kernel get out of sync the
449 ABI may break, preventing you from being able to add any rules.
450 This can adversely affect the booting process.
455 to temporarily disable the firewall to regain access to the network,
456 allowing you to fix the problem.
458 A packet is checked against the active ruleset in multiple places
459 in the protocol stack, under control of several sysctl variables.
460 These places and variables are shown below, and it is important to
461 have this picture in mind in order to design a correct ruleset.
462 .Bd -literal -offset indent
465 +----------->-----------+
467 [ip(6)_input] [ip(6)_output] net.inet(6).ip(6).fw.enable=1
470 [ether_demux] [ether_output_frame] net.link.ether.ipfw=1
472 +-->--[bdg_forward]-->--+ net.link.bridge.ipfw=1
478 times the same packet goes through the firewall can
479 vary between 0 and 4 depending on packet source and
480 destination, and system configuration.
482 Note that as packets flow through the stack, headers can be
483 stripped or added to it, and so they may or may not be available
485 E.g., incoming packets will include the MAC header when
489 but the same packets will have the MAC header stripped off when
496 Also note that each packet is always checked against the complete ruleset,
497 irrespective of the place where the check occurs, or the source of the packet.
498 If a rule contains some match patterns or actions which are not valid
499 for the place of invocation (e.g.\& trying to match a MAC header within
503 the match pattern will not match, but a
505 operator in front of such patterns
509 match on those packets.
510 It is thus the responsibility of
511 the programmer, if necessary, to write a suitable ruleset to
512 differentiate among the possible places.
514 rules can be useful here, as an example:
515 .Bd -literal -offset indent
516 # packets from ether_demux or bdg_forward
517 ipfw add 10 skipto 1000 all from any to any layer2 in
518 # packets from ip_input
519 ipfw add 10 skipto 2000 all from any to any not layer2 in
520 # packets from ip_output
521 ipfw add 10 skipto 3000 all from any to any not layer2 out
522 # packets from ether_output_frame
523 ipfw add 10 skipto 4000 all from any to any layer2 out
526 (yes, at the moment there is no way to differentiate between
527 ether_demux and bdg_forward).
529 Also note that only actions
538 frames and all other actions act as if they were
541 Full set of actions is supported for IP packets without
546 action does not divert
550 In general, each keyword or argument must be provided as
551 a separate command line argument, with no leading or trailing
553 Keywords are case-sensitive, whereas arguments may
554 or may not be case-sensitive depending on their nature
555 (e.g.\& uid's are, hostnames are not).
557 Some arguments (e.g., port or address lists) are comma-separated
559 In this case, spaces after commas ',' are allowed to make
560 the line more readable.
561 You can also put the entire
562 command (including flags) into a single argument.
563 E.g., the following forms are equivalent:
564 .Bd -literal -offset indent
565 ipfw -q add deny src-ip 10.0.0.0/24,127.0.0.1/8
566 ipfw -q add deny src-ip 10.0.0.0/24, 127.0.0.1/8
567 ipfw "-q add deny src-ip 10.0.0.0/24, 127.0.0.1/8"
570 The format of firewall rules is the following:
571 .Bd -ragged -offset indent
574 .Op Cm set Ar set_number
575 .Op Cm prob Ar match_probability
577 .Op Cm log Op Cm logamount Ar number
587 where the body of the rule specifies which information is used
588 for filtering packets, among the following:
590 .Bl -tag -width "Source and dest. addresses and ports" -offset XXX -compact
591 .It Layer-2 header fields
593 .It IPv4 and IPv6 Protocol
594 SCTP, TCP, UDP, ICMP, etc.
595 .It Source and dest. addresses and ports
599 .It Transmit and receive interface
601 .It Misc. IP header fields
602 Version, type of service, datagram length, identification,
606 .It IPv6 Extension headers
607 Fragmentation, Hop-by-Hop options,
608 Routing Headers, Source routing rthdr0, Mobile IPv6 rthdr2, IPSec options.
610 .It Misc. TCP header fields
611 TCP flags (SYN, FIN, ACK, RST, etc.),
612 sequence number, acknowledgment number,
620 When the packet can be associated with a local socket.
622 Whether a packet came from a divert socket (e.g.,
624 .It Fib annotation state
625 Whether a packet has been tagged for using a specific FIB (routing table)
626 in future forwarding decisions.
629 Note that some of the above information, e.g.\& source MAC or IP addresses and
630 TCP/UDP ports, can be easily spoofed, so filtering on those fields
631 alone might not guarantee the desired results.
632 .Bl -tag -width indent
634 Each rule is associated with a
636 in the range 1..65535, with the latter reserved for the
639 Rules are checked sequentially by rule number.
640 Multiple rules can have the same number, in which case they are
641 checked (and listed) according to the order in which they have
643 If a rule is entered without specifying a number, the kernel will
644 assign one in such a way that the rule becomes the last one
648 Automatic rule numbers are assigned by incrementing the last
649 non-default rule number by the value of the sysctl variable
650 .Ar net.inet.ip.fw.autoinc_step
651 which defaults to 100.
652 If this is not possible (e.g.\& because we would go beyond the
653 maximum allowed rule number), the number of the last
654 non-default value is used instead.
655 .It Cm set Ar set_number
656 Each rule is associated with a
659 Sets can be individually disabled and enabled, so this parameter
660 is of fundamental importance for atomic ruleset manipulation.
661 It can be also used to simplify deletion of groups of rules.
662 If a rule is entered without specifying a set number,
665 Set 31 is special in that it cannot be disabled,
666 and rules in set 31 are not deleted by the
668 command (but you can delete them with the
669 .Nm ipfw delete set 31
671 Set 31 is also used for the
674 .It Cm prob Ar match_probability
675 A match is only declared with the specified probability
676 (floating point number between 0 and 1).
677 This can be useful for a number of applications such as
678 random packet drop or
681 to simulate the effect of multiple paths leading to out-of-order
684 Note: this condition is checked before any other condition, including
691 .It Cm log Op Cm logamount Ar number
692 Packets matching a rule with the
694 keyword will be made available for logging in two ways:
695 if the sysctl variable
696 .Va net.inet.ip.fw.verbose
697 is set to 0 (default), one can use
702 This pseudo interface can be created manually after a system
703 boot by using the following command:
704 .Bd -literal -offset indent
705 # ifconfig ipfw0 create
708 Or, automatically at boot time by adding the following
712 .Bd -literal -offset indent
716 There is zero overhead when no
718 is attached to the pseudo interface.
721 .Va net.inet.ip.fw.verbose
722 is set to 1, packets will be logged to
726 facility up to a maximum of
731 is specified, the limit is taken from the sysctl variable
732 .Va net.inet.ip.fw.verbose_limit .
733 In both cases, a value of 0 means unlimited logging.
735 Once the limit is reached, logging can be re-enabled by
736 clearing the logging counter or the packet counter for that entry, see the
740 Note: logging is done after all other packet matching conditions
741 have been successfully verified, and before performing the final
742 action (accept, deny, etc.) on the packet.
744 When a packet matches a rule with the
746 keyword, the numeric tag for the given
748 in the range 1..65534 will be attached to the packet.
749 The tag acts as an internal marker (it is not sent out over
750 the wire) that can be used to identify these packets later on.
751 This can be used, for example, to provide trust between interfaces
752 and to start doing policy-based filtering.
753 A packet can have multiple tags at the same time.
754 Tags are "sticky", meaning once a tag is applied to a packet by a
755 matching rule it exists until explicit removal.
756 Tags are kept with the packet everywhere within the kernel, but are
757 lost when packet leaves the kernel, for example, on transmitting
758 packet out to the network or sending packet to a
762 To check for previously applied tags, use the
765 To delete previously applied tag, use the
769 Note: since tags are kept with the packet everywhere in kernelspace,
770 they can be set and unset anywhere in the kernel network subsystem
773 facility), not only by means of the
779 For example, there can be a specialized
781 node doing traffic analyzing and tagging for later inspecting
783 .It Cm untag Ar number
784 When a packet matches a rule with the
786 keyword, the tag with the number
788 is searched among the tags attached to this packet and,
789 if found, removed from it.
790 Other tags bound to packet, if present, are left untouched.
792 When a packet matches a rule with the
794 keyword, the ALTQ identifier for the given
799 Note that this ALTQ tag is only meaningful for packets going "out" of IPFW,
800 and not being rejected or going to divert sockets.
801 Note that if there is insufficient memory at the time the packet is
802 processed, it will not be tagged, so it is wise to make your ALTQ
803 "default" queue policy account for this.
806 rules match a single packet, only the first one adds the ALTQ classification
808 In doing so, traffic may be shaped by using
809 .Cm count Cm altq Ar queue
810 rules for classification early in the ruleset, then later applying
811 the filtering decision.
816 rules may come later and provide the actual filtering decisions in
817 addition to the fallback ALTQ tag.
821 to set up the queues before IPFW will be able to look them up by name,
822 and if the ALTQ disciplines are rearranged, the rules in containing the
823 queue identifiers in the kernel will likely have gone stale and need
825 Stale queue identifiers will probably result in misclassification.
827 All system ALTQ processing can be turned on or off via
832 .Cm disable Ar altq .
834 .Va net.inet.ip.fw.one_pass
835 is irrelevant to ALTQ traffic shaping, as the actual rule action is followed
836 always after adding an ALTQ tag.
839 A rule can be associated with one of the following actions, which
840 will be executed when the packet matches the body of the rule.
841 .Bl -tag -width indent
842 .It Cm allow | accept | pass | permit
843 Allow packets that match rule.
844 The search terminates.
845 .It Cm check-state Op Ar :flowname | Cm :any
846 Checks the packet against the dynamic ruleset.
847 If a match is found, execute the action associated with
848 the rule which generated this dynamic rule, otherwise
849 move to the next rule.
852 rules do not have a body.
855 rule is found, the dynamic ruleset is checked at the first
862 is symbolic name assigned to dynamic rule by
867 can be used to ignore states flowname when matching.
870 keyword is special name used for compatibility with old rulesets.
872 Update counters for all packets that match rule.
873 The search continues with the next rule.
875 Discard packets that match this rule.
876 The search terminates.
877 .It Cm divert Ar port
878 Divert packets that match this rule to the
882 The search terminates.
883 .It Cm fwd | forward Ar ipaddr | tablearg Ns Op , Ns Ar port
884 Change the next-hop on matching packets to
886 which can be an IP address or a host name.
887 The next hop can also be supplied by the last table
888 looked up for the packet by using the
890 keyword instead of an explicit address.
891 The search terminates if this rule matches.
895 is a local address, then matching packets will be forwarded to
897 (or the port number in the packet if one is not specified in the rule)
898 on the local machine.
902 is not a local address, then the port number
903 (if specified) is ignored, and the packet will be
904 forwarded to the remote address, using the route as found in
905 the local routing table for that IP.
909 rule will not match layer-2 packets (those received
910 on ether_input, ether_output, or bridged).
914 action does not change the contents of the packet at all.
915 In particular, the destination address remains unmodified, so
916 packets forwarded to another system will usually be rejected by that system
917 unless there is a matching rule on that system to capture them.
918 For packets forwarded locally,
919 the local address of the socket will be
920 set to the original destination address of the packet.
923 entry look rather weird but is intended for
924 use with transparent proxy servers.
925 .It Cm nat Ar nat_nr | tablearg
928 (for network address translation, address redirect, etc.):
930 .Sx NETWORK ADDRESS TRANSLATION (NAT)
931 Section for further information.
932 .It Cm nat64lsn Ar name
933 Pass packet to a stateful NAT64 instance (for IPv6/IPv4 network address and
934 protocol translation): see the
935 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
936 Section for further information.
937 .It Cm nat64stl Ar name
938 Pass packet to a stateless NAT64 instance (for IPv6/IPv4 network address and
939 protocol translation): see the
940 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
941 Section for further information.
942 .It Cm nat64clat Ar name
943 Pass packet to a CLAT NAT64 instance (for client-side IPv6/IPv4 network address and
944 protocol translation): see the
945 .Sx IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
946 Section for further information.
948 Pass packet to a NPTv6 instance (for IPv6-to-IPv6 network prefix translation):
950 .Sx IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
951 Section for further information.
952 .It Cm pipe Ar pipe_nr
956 (for bandwidth limitation, delay, etc.).
958 .Sx TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
959 Section for further information.
960 The search terminates; however, on exit from the pipe and if
964 .Va net.inet.ip.fw.one_pass
965 is not set, the packet is passed again to the firewall code
966 starting from the next rule.
967 .It Cm queue Ar queue_nr
971 (for bandwidth limitation using WF2Q+).
977 Discard packets that match this rule, and if the
978 packet is a TCP packet, try to send a TCP reset (RST) notice.
979 The search terminates.
981 Discard packets that match this rule, and if the
982 packet is a TCP packet, try to send a TCP reset (RST) notice.
983 The search terminates.
984 .It Cm skipto Ar number | tablearg
985 Skip all subsequent rules numbered less than
987 The search continues with the first rule numbered
990 It is possible to use the
992 keyword with a skipto for a
995 Skipto may work either in O(log(N)) or in O(1) depending
996 on amount of memory and/or sysctl variables.
999 section for more details.
1000 .It Cm call Ar number | tablearg
1001 The current rule number is saved in the internal stack and
1002 ruleset processing continues with the first rule numbered
1005 If later a rule with the
1007 action is encountered, the processing returns to the first rule
1010 rule plus one or higher
1011 (the same behaviour as with packets returning from
1016 This could be used to make somewhat like an assembly language
1018 calls to rules with common checks for different interfaces, etc.
1020 Rule with any number could be called, not just forward jumps as with
1022 So, to prevent endless loops in case of mistakes, both
1026 actions don't do any jumps and simply go to the next rule if memory
1027 cannot be allocated or stack overflowed/underflowed.
1029 Internally stack for rule numbers is implemented using
1031 facility and currently has size of 16 entries.
1032 As mbuf tags are lost when packet leaves the kernel,
1034 should not be used in subroutines to avoid endless loops
1035 and other undesired effects.
1037 Takes rule number saved to internal stack by the last
1039 action and returns ruleset processing to the first rule
1040 with number greater than number of corresponding
1043 See description of the
1045 action for more details.
1051 and thus are unconditional, but
1053 command-line utility currently requires every action except
1056 While it is sometimes useful to return only on some packets,
1057 usually you want to print just
1060 A workaround for this is to use new syntax and
1063 .Bd -literal -offset indent
1064 # Add a rule without actual body
1065 ipfw add 2999 return via any
1067 # List rules without "from any to any" part
1071 This cosmetic annoyance may be fixed in future releases.
1073 Send a copy of packets matching this rule to the
1075 socket bound to port
1077 The search continues with the next rule.
1078 .It Cm unreach Ar code
1079 Discard packets that match this rule, and try to send an ICMP
1080 unreachable notice with code
1084 is a number from 0 to 255, or one of these aliases:
1085 .Cm net , host , protocol , port ,
1086 .Cm needfrag , srcfail , net-unknown , host-unknown ,
1087 .Cm isolated , net-prohib , host-prohib , tosnet ,
1088 .Cm toshost , filter-prohib , host-precedence
1090 .Cm precedence-cutoff .
1091 The search terminates.
1092 .It Cm unreach6 Ar code
1093 Discard packets that match this rule, and try to send an ICMPv6
1094 unreachable notice with code
1098 is a number from 0, 1, 3 or 4, or one of these aliases:
1099 .Cm no-route, admin-prohib, address
1102 The search terminates.
1103 .It Cm netgraph Ar cookie
1104 Divert packet into netgraph with given
1106 The search terminates.
1107 If packet is later returned from netgraph it is either
1108 accepted or continues with the next rule, depending on
1109 .Va net.inet.ip.fw.one_pass
1111 .It Cm ngtee Ar cookie
1112 A copy of packet is diverted into netgraph, original
1113 packet continues with the next rule.
1116 for more information on
1121 .It Cm setfib Ar fibnum | tablearg
1122 The packet is tagged so as to use the FIB (routing table)
1124 in any subsequent forwarding decisions.
1125 In the current implementation, this is limited to the values 0 through 15, see
1127 Processing continues at the next rule.
1128 It is possible to use the
1130 keyword with setfib.
1131 If the tablearg value is not within the compiled range of fibs,
1132 the packet's fib is set to 0.
1133 .It Cm setdscp Ar DSCP | number | tablearg
1134 Set specified DiffServ codepoint for an IPv4/IPv6 packet.
1135 Processing continues at the next rule.
1136 Supported values are:
1182 Additionally, DSCP value can be specified by number (0..63).
1183 It is also possible to use the
1185 keyword with setdscp.
1186 If the tablearg value is not within the 0..63 range, lower 6 bits of supplied
1188 .It Cm tcp-setmss Ar mss
1189 Set the Maximum Segment Size (MSS) in the TCP segment to value
1193 should be loaded or kernel should have
1194 .Cm options IPFIREWALL_PMOD
1195 to be able use this action.
1196 This command does not change a packet if original MSS value is lower than
1198 Both TCP over IPv4 and over IPv6 are supported.
1199 Regardless of matched a packet or not by the
1201 rule, the search continues with the next rule.
1203 Queue and reassemble IPv4 fragments.
1204 If the packet is not fragmented, counters are updated and
1205 processing continues with the next rule.
1206 If the packet is the last logical fragment, the packet is reassembled and, if
1207 .Va net.inet.ip.fw.one_pass
1208 is set to 0, processing continues with the next rule.
1209 Otherwise, the packet is allowed to pass and the search terminates.
1210 If the packet is a fragment in the middle of a logical group of fragments,
1212 processing stops immediately.
1214 Fragment handling can be tuned via
1215 .Va net.inet.ip.maxfragpackets
1217 .Va net.inet.ip.maxfragsperpacket
1218 which limit, respectively, the maximum number of processable
1219 fragments (default: 800) and
1220 the maximum number of fragments per packet (default: 16).
1222 NOTA BENE: since fragments do not contain port numbers,
1223 they should be avoided with the
1226 Alternatively, direction-based (like
1230 ) and source-based (like
1232 ) match patterns can be used to select fragments.
1234 Usually a simple rule like:
1235 .Bd -literal -offset indent
1236 # reassemble incoming fragments
1237 ipfw add reass all from any to any in
1240 is all you need at the beginning of your ruleset.
1242 Discard packets that match this rule, and if the packet is an SCTP packet,
1243 try to send an SCTP packet containing an ABORT chunk.
1244 The search terminates.
1246 Discard packets that match this rule, and if the packet is an SCTP packet,
1247 try to send an SCTP packet containing an ABORT chunk.
1248 The search terminates.
1251 The body of a rule contains zero or more patterns (such as
1252 specific source and destination addresses or ports,
1253 protocol options, incoming or outgoing interfaces, etc.)
1254 that the packet must match in order to be recognised.
1255 In general, the patterns are connected by (implicit)
1257 operators -- i.e., all must match in order for the
1259 Individual patterns can be prefixed by the
1261 operator to reverse the result of the match, as in
1263 .Dl "ipfw add 100 allow ip from not 1.2.3.4 to any"
1265 Additionally, sets of alternative match patterns
1267 can be constructed by putting the patterns in
1268 lists enclosed between parentheses ( ) or braces { }, and
1271 operator as follows:
1273 .Dl "ipfw add 100 allow ip from { x or not y or z } to any"
1275 Only one level of parentheses is allowed.
1276 Beware that most shells have special meanings for parentheses
1277 or braces, so it is advisable to put a backslash \\ in front of them
1278 to prevent such interpretations.
1280 The body of a rule must in general include a source and destination
1284 can be used in various places to specify that the content of
1285 a required field is irrelevant.
1287 The rule body has the following format:
1288 .Bd -ragged -offset indent
1289 .Op Ar proto Cm from Ar src Cm to Ar dst
1293 The first part (proto from src to dst) is for backward
1294 compatibility with earlier versions of
1298 any match pattern (including MAC headers, IP protocols,
1299 addresses and ports) can be specified in the
1303 Rule fields have the following meaning:
1304 .Bl -tag -width indent
1305 .It Ar proto : protocol | Cm { Ar protocol Cm or ... }
1306 .It Ar protocol : Oo Cm not Oc Ar protocol-name | protocol-number
1307 An IP protocol specified by number or name
1308 (for a complete list see
1309 .Pa /etc/protocols ) ,
1310 or one of the following keywords:
1311 .Bl -tag -width indent
1313 Matches IPv4 packets.
1315 Matches IPv6 packets.
1324 option will be treated as inner protocol.
1332 .Cm { Ar protocol Cm or ... }
1335 is provided for convenience only but its use is deprecated.
1336 .It Ar src No and Ar dst : Bro Cm addr | Cm { Ar addr Cm or ... } Brc Op Oo Cm not Oc Ar ports
1337 An address (or a list, see below)
1338 optionally followed by
1344 with multiple addresses) is provided for convenience only and
1345 its use is discouraged.
1346 .It Ar addr : Oo Cm not Oc Bro
1347 .Cm any | me | me6 |
1348 .Cm table Ns Pq Ar name Ns Op , Ns Ar value
1349 .Ar | addr-list | addr-set
1351 .Bl -tag -width indent
1353 Matches any IP address.
1355 Matches any IP address configured on an interface in the system.
1357 Matches any IPv6 address configured on an interface in the system.
1358 The address list is evaluated at the time the packet is
1360 .It Cm table Ns Pq Ar name Ns Op , Ns Ar value
1361 Matches any IPv4 or IPv6 address for which an entry exists in the lookup table
1363 If an optional 32-bit unsigned
1365 is also specified, an entry will match only if it has this value.
1368 section below for more information on lookup tables.
1370 .It Ar addr-list : ip-addr Ns Op Ns , Ns Ar addr-list
1372 A host or subnet address specified in one of the following ways:
1373 .Bl -tag -width indent
1374 .It Ar numeric-ip | hostname
1375 Matches a single IPv4 address, specified as dotted-quad or a hostname.
1376 Hostnames are resolved at the time the rule is added to the firewall list.
1377 .It Ar addr Ns / Ns Ar masklen
1378 Matches all addresses with base
1380 (specified as an IP address, a network number, or a hostname)
1384 As an example, 1.2.3.4/25 or 1.2.3.0/25 will match
1385 all IP numbers from 1.2.3.0 to 1.2.3.127 .
1386 .It Ar addr Ns : Ns Ar mask
1387 Matches all addresses with base
1389 (specified as an IP address, a network number, or a hostname)
1392 specified as a dotted quad.
1393 As an example, 1.2.3.4:255.0.255.0 or 1.0.3.0:255.0.255.0 will match
1395 This form is advised only for non-contiguous
1397 It is better to resort to the
1398 .Ar addr Ns / Ns Ar masklen
1399 format for contiguous masks, which is more compact and less
1402 .It Ar addr-set : addr Ns Oo Ns / Ns Ar masklen Oc Ns Cm { Ns Ar list Ns Cm }
1403 .It Ar list : Bro Ar num | num-num Brc Ns Op Ns , Ns Ar list
1404 Matches all addresses with base address
1406 (specified as an IP address, a network number, or a hostname)
1407 and whose last byte is in the list between braces { } .
1408 Note that there must be no spaces between braces and
1409 numbers (spaces after commas are allowed).
1410 Elements of the list can be specified as single entries
1414 field is used to limit the size of the set of addresses,
1415 and can have any value between 24 and 32.
1417 it will be assumed as 24.
1419 This format is particularly useful to handle sparse address sets
1420 within a single rule.
1421 Because the matching occurs using a
1422 bitmask, it takes constant time and dramatically reduces
1423 the complexity of rulesets.
1425 As an example, an address specified as 1.2.3.4/24{128,35-55,89}
1426 or 1.2.3.0/24{128,35-55,89}
1427 will match the following IP addresses:
1429 1.2.3.128, 1.2.3.35 to 1.2.3.55, 1.2.3.89 .
1430 .It Ar addr6-list : ip6-addr Ns Op Ns , Ns Ar addr6-list
1432 A host or subnet specified one of the following ways:
1433 .Bl -tag -width indent
1434 .It Ar numeric-ip | hostname
1435 Matches a single IPv6 address as allowed by
1438 Hostnames are resolved at the time the rule is added to the firewall
1440 .It Ar addr Ns / Ns Ar masklen
1441 Matches all IPv6 addresses with base
1443 (specified as allowed by
1449 .It Ar addr Ns / Ns Ar mask
1450 Matches all IPv6 addresses with base
1452 (specified as allowed by
1457 specified as allowed by
1459 As an example, fe::640:0:0/ffff::ffff:ffff:0:0 will match
1461 This form is advised only for non-contiguous
1463 It is better to resort to the
1464 .Ar addr Ns / Ns Ar masklen
1465 format for contiguous masks, which is more compact and less
1469 No support for sets of IPv6 addresses is provided because IPv6 addresses
1470 are typically random past the initial prefix.
1471 .It Ar ports : Bro Ar port | port Ns \&- Ns Ar port Ns Brc Ns Op , Ns Ar ports
1472 For protocols which support port numbers (such as SCTP, TCP and UDP), optional
1474 may be specified as one or more ports or port ranges, separated
1475 by commas but no spaces, and an optional
1480 notation specifies a range of ports (including boundaries).
1484 may be used instead of numeric port values.
1485 The length of the port list is limited to 30 ports or ranges,
1486 though one can specify larger ranges by using an
1490 section of the rule.
1494 can be used to escape the dash
1496 character in a service name (from a shell, the backslash must be
1497 typed twice to avoid the shell itself interpreting it as an escape
1500 .Dl "ipfw add count tcp from any ftp\e\e-data-ftp to any"
1502 Fragmented packets which have a non-zero offset (i.e., not the first
1503 fragment) will never match a rule which has one or more port
1507 option for details on matching fragmented packets.
1509 .Ss RULE OPTIONS (MATCH PATTERNS)
1510 Additional match patterns can be used within
1512 Zero or more of these so-called
1514 can be present in a rule, optionally prefixed by the
1516 operand, and possibly grouped into
1519 The following match patterns can be used (listed in alphabetical order):
1520 .Bl -tag -width indent
1521 .It Cm // this is a comment.
1522 Inserts the specified text as a comment in the rule.
1523 Everything following // is considered as a comment and stored in the rule.
1524 You can have comment-only rules, which are listed as having a
1526 action followed by the comment.
1530 .It Cm defer-immediate-action | defer-action
1531 A rule with this option will not perform normal action
1533 This option is intended to be used with
1537 as the dynamic rule, created but ignored on match, will work
1542 .Cm defer-immediate-action
1543 create a dynamic rule and continue with the next rule without actually
1544 performing the action part of this rule.
1545 When the rule is later activated via the state table, the action is
1548 Matches only packets generated by a divert socket.
1549 .It Cm diverted-loopback
1550 Matches only packets coming from a divert socket back into the IP stack
1552 .It Cm diverted-output
1553 Matches only packets going from a divert socket back outward to the IP
1554 stack output for delivery.
1555 .It Cm dst-ip Ar ip-address
1556 Matches IPv4 packets whose destination IP is one of the address(es)
1557 specified as argument.
1558 .It Bro Cm dst-ip6 | dst-ipv6 Brc Ar ip6-address
1559 Matches IPv6 packets whose destination IP is one of the address(es)
1560 specified as argument.
1561 .It Cm dst-port Ar ports
1562 Matches IP packets whose destination port is one of the port(s)
1563 specified as argument.
1565 Matches TCP packets that have the RST or ACK bits set.
1566 .It Cm ext6hdr Ar header
1567 Matches IPv6 packets containing the extended header given by
1569 Supported headers are:
1575 any type of Routing Header
1577 Source routing Routing Header Type 0
1579 Mobile IPv6 Routing Header Type 2
1583 IPSec authentication headers
1585 and IPsec encapsulated security payload headers
1587 .It Cm fib Ar fibnum
1588 Matches a packet that has been tagged to use
1589 the given FIB (routing table) number.
1590 .It Cm flow Ar table Ns Pq Ar name Ns Op , Ns Ar value
1591 Search for the flow entry in lookup table
1593 If not found, the match fails.
1594 Otherwise, the match succeeds and
1596 is set to the value extracted from the table.
1598 This option can be useful to quickly dispatch traffic based on
1599 certain packet fields.
1602 section below for more information on lookup tables.
1603 .It Cm flow-id Ar labels
1604 Matches IPv6 packets containing any of the flow labels given in
1607 is a comma separated list of numeric flow labels.
1609 Matches IPv4 packets whose
1611 field contains the comma separated list of IPv4 fragmentation
1612 options specified in
1614 The recognized options are:
1616 .Pq Dv don't fragment ,
1618 .Pq Dv more fragments ,
1620 .Pq Dv reserved fragment bit
1622 .Pq Dv non-zero fragment offset .
1623 The absence of a particular options may be denoted
1627 Empty list of options defaults to matching on non-zero fragment offset.
1628 Such rule would match all not the first fragment datagrams,
1630 This is a backward compatibility with older rulesets.
1632 Matches all TCP or UDP packets sent by or received for a
1636 may be specified by name or number.
1638 Matches all TCP or UDP packets sent by or received for the
1639 jail whose ID or name is
1641 .It Cm icmptypes Ar types
1642 Matches ICMP packets whose ICMP type is in the list
1644 The list may be specified as any combination of
1645 individual types (numeric) separated by commas.
1646 .Em Ranges are not allowed .
1647 The supported ICMP types are:
1651 destination unreachable
1659 router advertisement
1663 time-to-live exceeded
1675 address mask request
1677 and address mask reply
1679 .It Cm icmp6types Ar types
1680 Matches ICMP6 packets whose ICMP6 type is in the list of
1682 The list may be specified as any combination of
1683 individual types (numeric) separated by commas.
1684 .Em Ranges are not allowed .
1686 Matches incoming or outgoing packets, respectively.
1690 are mutually exclusive (in fact,
1694 .It Cm ipid Ar id-list
1695 Matches IPv4 packets whose
1697 field has value included in
1699 which is either a single value or a list of values or ranges
1700 specified in the same way as
1702 .It Cm iplen Ar len-list
1703 Matches IP packets whose total length, including header and data, is
1706 which is either a single value or a list of values or ranges
1707 specified in the same way as
1709 .It Cm ipoptions Ar spec
1710 Matches packets whose IPv4 header contains the comma separated list of
1711 options specified in
1713 The supported IP options are:
1716 (strict source route),
1718 (loose source route),
1720 (record packet route) and
1723 The absence of a particular option may be denoted
1726 .It Cm ipprecedence Ar precedence
1727 Matches IPv4 packets whose precedence field is equal to
1730 Matches packets that have IPSEC history associated with them
1731 (i.e., the packet comes encapsulated in IPSEC, the kernel
1732 has IPSEC support, and can correctly decapsulate it).
1734 Note that specifying
1736 is different from specifying
1738 as the latter will only look at the specific IP protocol field,
1739 irrespective of IPSEC kernel support and the validity of the IPSEC data.
1741 Further note that this flag is silently ignored in kernels without
1743 It does not affect rule processing when given and the
1744 rules are handled as if with no
1747 .It Cm iptos Ar spec
1748 Matches IPv4 packets whose
1750 field contains the comma separated list of
1751 service types specified in
1753 The supported IP types of service are:
1756 .Pq Dv IPTOS_LOWDELAY ,
1758 .Pq Dv IPTOS_THROUGHPUT ,
1760 .Pq Dv IPTOS_RELIABILITY ,
1762 .Pq Dv IPTOS_MINCOST ,
1764 .Pq Dv IPTOS_ECN_CE .
1765 The absence of a particular type may be denoted
1768 .It Cm dscp spec Ns Op , Ns Ar spec
1769 Matches IPv4/IPv6 packets whose
1771 field value is contained in
1774 Multiple values can be specified via
1775 the comma separated list.
1776 Value can be one of keywords used in
1778 action or exact number.
1779 .It Cm ipttl Ar ttl-list
1780 Matches IPv4 packets whose time to live is included in
1782 which is either a single value or a list of values or ranges
1783 specified in the same way as
1785 .It Cm ipversion Ar ver
1786 Matches IP packets whose IP version field is
1788 .It Cm keep-state Op Ar :flowname
1789 Upon a match, the firewall will create a dynamic rule, whose
1790 default behaviour is to match bidirectional traffic between
1791 source and destination IP/port using the same protocol.
1792 The rule has a limited lifetime (controlled by a set of
1794 variables), and the lifetime is refreshed every time a matching
1798 is used to assign additional to addresses, ports and protocol parameter
1800 It can be used for more accurate matching by
1805 keyword is special name used for compatibility with old rulesets.
1807 Matches only layer2 packets, i.e., those passed to
1809 from ether_demux() and ether_output_frame().
1810 .It Cm limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N Op Ar :flowname
1811 The firewall will only allow
1813 connections with the same
1814 set of parameters as specified in the rule.
1816 of source and destination addresses and ports can be
1818 .It Cm lookup Bro Cm dst-ip | dst-port | src-ip | src-port | uid | jail Brc Ar name
1819 Search an entry in lookup table
1821 that matches the field specified as argument.
1822 If not found, the match fails.
1823 Otherwise, the match succeeds and
1825 is set to the value extracted from the table.
1827 This option can be useful to quickly dispatch traffic based on
1828 certain packet fields.
1831 section below for more information on lookup tables.
1832 .It Cm { MAC | mac } Ar dst-mac src-mac
1833 Match packets with a given
1837 addresses, specified as the
1839 keyword (matching any MAC address), or six groups of hex digits
1840 separated by colons,
1841 and optionally followed by a mask indicating the significant bits.
1842 The mask may be specified using either of the following methods:
1843 .Bl -enum -width indent
1847 followed by the number of significant bits.
1848 For example, an address with 33 significant bits could be specified as:
1850 .Dl "MAC 10:20:30:40:50:60/33 any"
1854 followed by a bitmask specified as six groups of hex digits separated
1856 For example, an address in which the last 16 bits are significant could
1859 .Dl "MAC 10:20:30:40:50:60&00:00:00:00:ff:ff any"
1861 Note that the ampersand character has a special meaning in many shells
1862 and should generally be escaped.
1864 Note that the order of MAC addresses (destination first,
1866 the same as on the wire, but the opposite of the one used for
1868 .It Cm mac-type Ar mac-type
1869 Matches packets whose Ethernet Type field
1870 corresponds to one of those specified as argument.
1872 is specified in the same way as
1874 (i.e., one or more comma-separated single values or ranges).
1875 You can use symbolic names for known values such as
1876 .Em vlan , ipv4, ipv6 .
1877 Values can be entered as decimal or hexadecimal (if prefixed by 0x),
1878 and they are always printed as hexadecimal (unless the
1880 option is used, in which case symbolic resolution will be attempted).
1881 .It Cm proto Ar protocol
1882 Matches packets with the corresponding IP protocol.
1884 Upon a match, the firewall will create a dynamic rule as if
1887 However, this option doesn't imply an implicit
1891 .It Cm recv | xmit | via Brq Ar ifX | Ar if Ns Cm * | Ar table Ns Po Ar name Ns Oo , Ns Ar value Oc Pc | Ar ipno | Ar any
1892 Matches packets received, transmitted or going through,
1893 respectively, the interface specified by exact name
1897 by IP address, or through some interface.
1900 may be used to match interface by its kernel ifindex.
1903 section below for more information on lookup tables.
1907 keyword causes the interface to always be checked.
1914 then only the receive or transmit interface (respectively)
1916 By specifying both, it is possible to match packets based on
1917 both receive and transmit interface, e.g.:
1919 .Dl "ipfw add deny ip from any to any out recv ed0 xmit ed1"
1923 interface can be tested on either incoming or outgoing packets,
1926 interface can only be tested on outgoing packets.
1931 is invalid) whenever
1935 A packet might not have a receive or transmit interface: packets
1936 originating from the local host have no receive interface,
1937 while packets destined for the local host have no transmit
1939 .It Cm set-limit Bro Cm src-addr | src-port | dst-addr | dst-port Brc Ar N
1942 but does not have an implicit
1946 Matches TCP packets that have the SYN bit set but no ACK bit.
1947 This is the short form of
1948 .Dq Li tcpflags\ syn,!ack .
1950 Matches packets that are associated to a local socket and
1951 for which the SO_USER_COOKIE socket option has been set
1952 to a non-zero value.
1953 As a side effect, the value of the
1954 option is made available as
1956 value, which in turn can be used as
1961 .It Cm src-ip Ar ip-address
1962 Matches IPv4 packets whose source IP is one of the address(es)
1963 specified as an argument.
1964 .It Cm src-ip6 Ar ip6-address
1965 Matches IPv6 packets whose source IP is one of the address(es)
1966 specified as an argument.
1967 .It Cm src-port Ar ports
1968 Matches IP packets whose source port is one of the port(s)
1969 specified as argument.
1970 .It Cm tagged Ar tag-list
1971 Matches packets whose tags are included in
1973 which is either a single value or a list of values or ranges
1974 specified in the same way as
1976 Tags can be applied to the packet using
1978 rule action parameter (see it's description for details on tags).
1979 .It Cm tcpack Ar ack
1981 Match if the TCP header acknowledgment number field is set to
1983 .It Cm tcpdatalen Ar tcpdatalen-list
1984 Matches TCP packets whose length of TCP data is
1985 .Ar tcpdatalen-list ,
1986 which is either a single value or a list of values or ranges
1987 specified in the same way as
1989 .It Cm tcpflags Ar spec
1991 Match if the TCP header contains the comma separated list of
1994 The supported TCP flags are:
2003 The absence of a particular flag may be denoted
2006 A rule which contains a
2008 specification can never match a fragmented packet which has
2012 option for details on matching fragmented packets.
2013 .It Cm tcpmss Ar tcpmss-list
2014 Matches TCP packets whose MSS (maximum segment size) value is set to
2016 which is either a single value or a list of values or ranges
2017 specified in the same way as
2019 .It Cm tcpseq Ar seq
2021 Match if the TCP header sequence number field is set to
2023 .It Cm tcpwin Ar tcpwin-list
2024 Matches TCP packets whose header window field is set to
2026 which is either a single value or a list of values or ranges
2027 specified in the same way as
2029 .It Cm tcpoptions Ar spec
2031 Match if the TCP header contains the comma separated list of
2032 options specified in
2034 The supported TCP options are:
2037 (maximum segment size),
2039 (tcp window advertisement),
2043 (rfc1323 timestamp) and
2045 (rfc1644 t/tcp connection count).
2046 The absence of a particular option may be denoted
2050 Match all TCP or UDP packets sent by or received for a
2054 may be matched by name or identification number.
2056 For incoming packets,
2057 a routing table lookup is done on the packet's source address.
2058 If the interface on which the packet entered the system matches the
2059 outgoing interface for the route,
2061 If the interfaces do not match up,
2062 the packet does not match.
2063 All outgoing packets or packets with no incoming interface match.
2065 The name and functionality of the option is intentionally similar to
2066 the Cisco IOS command:
2068 .Dl ip verify unicast reverse-path
2070 This option can be used to make anti-spoofing rules to reject all
2071 packets with source addresses not from this interface.
2075 For incoming packets,
2076 a routing table lookup is done on the packet's source address.
2077 If a route to the source address exists, but not the default route
2078 or a blackhole/reject route, the packet matches.
2079 Otherwise, the packet does not match.
2080 All outgoing packets match.
2082 The name and functionality of the option is intentionally similar to
2083 the Cisco IOS command:
2085 .Dl ip verify unicast source reachable-via any
2087 This option can be used to make anti-spoofing rules to reject all
2088 packets whose source address is unreachable.
2090 For incoming packets, the packet's source address is checked if it
2091 belongs to a directly connected network.
2092 If the network is directly connected, then the interface the packet
2093 came on in is compared to the interface the network is connected to.
2094 When incoming interface and directly connected interface are not the
2095 same, the packet does not match.
2096 Otherwise, the packet does match.
2097 All outgoing packets match.
2099 This option can be used to make anti-spoofing rules to reject all
2100 packets that pretend to be from a directly connected network but do
2101 not come in through that interface.
2102 This option is similar to but more restricted than
2104 because it engages only on packets with source addresses of directly
2105 connected networks instead of all source addresses.
2108 Lookup tables are useful to handle large sparse sets of
2109 addresses or other search keys (e.g., ports, jail IDs, interface names).
2110 In the rest of this section we will use the term ``key''.
2111 Table name needs to match the following spec:
2113 Tables with the same name can be created in different
2115 However, rule links to the tables in
2118 This behavior can be controlled by
2119 .Va net.inet.ip.fw.tables_sets
2123 section for more information.
2124 There may be up to 65535 different lookup tables.
2126 The following table types are supported:
2127 .Bl -tag -width indent
2128 .It Ar table-type : Ar addr | iface | number | flow
2129 .It Ar table-key : Ar addr Ns Oo / Ns Ar masklen Oc | iface-name | number | flow-spec
2130 .It Ar flow-spec : Ar flow-field Ns Op , Ns Ar flow-spec
2131 .It Ar flow-field : src-ip | proto | src-port | dst-ip | dst-port
2133 Matches IPv4 or IPv6 address.
2134 Each entry is represented by an
2135 .Ar addr Ns Op / Ns Ar masklen
2136 and will match all addresses with base
2138 (specified as an IPv4/IPv6 address, or a hostname) and mask width of
2143 is not specified, it defaults to 32 for IPv4 and 128 for IPv6.
2144 When looking up an IP address in a table, the most specific
2147 Matches interface names.
2148 Each entry is represented by string treated as interface name.
2149 Wildcards are not supported.
2151 Matches protocol ports, uids/gids or jail IDs.
2152 Each entry is represented by 32-bit unsigned integer.
2153 Ranges are not supported.
2155 Matches packet fields specified by
2157 type suboptions with table entries.
2160 Tables require explicit creation via
2164 The following creation options are supported:
2165 .Bl -tag -width indent
2166 .It Ar create-options : Ar create-option | create-options
2167 .It Ar create-option : Cm type Ar table-type | Cm valtype Ar value-mask | Cm algo Ar algo-desc |
2168 .Cm limit Ar number | Cm locked | Cm missing | Cm or-flush
2174 Table algorithm to use (see below).
2176 Maximum number of items that may be inserted into table.
2178 Restrict any table modifications.
2180 Do not fail if table already exists and has exactly same options as new one.
2182 Flush existing table with same name instead of returning error.
2185 so existing table must be compatible with new one.
2188 Some of these options may be modified later via
2191 The following options can be changed:
2192 .Bl -tag -width indent
2193 .It Ar modify-options : Ar modify-option | modify-options
2194 .It Ar modify-option : Cm limit Ar number
2196 Alter maximum number of items that may be inserted into table.
2199 Additionally, table can be locked or unlocked using
2207 can be swapped with each other using
2210 Swap may fail if tables limits are set and data exchange
2211 would result in limits hit.
2212 Operation is performed atomically.
2214 One or more entries can be added to a table at once using
2217 Addition of all items are performed atomically.
2218 By default, error in addition of one entry does not influence
2219 addition of other entries.
2220 However, non-zero error code is returned in that case.
2223 keyword may be specified before
2225 to indicate all-or-none add request.
2227 One or more entries can be removed from a table at once using
2230 By default, error in removal of one entry does not influence
2231 removing of other entries.
2232 However, non-zero error code is returned in that case.
2234 It may be possible to check what entry will be found on particular
2240 This functionality is optional and may be unsupported in some algorithms.
2242 The following operations can be performed on
2247 .Bl -tag -width indent
2251 Removes all entries.
2253 Shows generic table information.
2255 Shows generic table information and algo-specific data.
2258 The following lookup algorithms are supported:
2259 .Bl -tag -width indent
2260 .It Ar algo-desc : algo-name | "algo-name algo-data"
2261 .It Ar algo-name: Ar addr:radix | addr:hash | iface:array | number:array | flow:hash
2263 Separate Radix trees for IPv4 and IPv6, the same way as the routing table (see
2269 Separate auto-growing hashes for IPv4 and IPv6.
2270 Accepts entries with the same mask length specified initially via
2271 .Cm "addr:hash masks=/v4,/v6"
2272 algorithm creation options.
2273 Assume /32 and /128 masks by default.
2274 Search removes host bits (according to mask) from supplied address and checks
2275 resulting key in appropriate hash.
2276 Mostly optimized for /64 and byte-ranged IPv6 masks.
2278 Array storing sorted indexes for entries which are presented in the system.
2279 Optimized for very fast lookup.
2281 Array storing sorted u32 numbers.
2283 Auto-growing hash storing flow entries.
2284 Search calculates hash on required packet fields and searches for matching
2285 entries in selected bucket.
2290 feature provides the ability to use a value, looked up in the table, as
2291 the argument for a rule action, action parameter or rule option.
2292 This can significantly reduce number of rules in some configurations.
2293 If two tables are used in a rule, the result of the second (destination)
2296 Each record may hold one or more values according to
2298 This mask is set on table creation via
2301 The following value types are supported:
2302 .Bl -tag -width indent
2303 .It Ar value-mask : Ar value-type Ns Op , Ns Ar value-mask
2304 .It Ar value-type : Ar skipto | pipe | fib | nat | dscp | tag | divert |
2305 .Ar netgraph | limit | ipv4
2307 rule number to jump to.
2311 fib number to match/set.
2313 nat number to jump to.
2315 dscp value to match/set.
2317 tag number to match/set.
2319 port number to divert traffic to.
2321 hook number to move packet to.
2323 maximum number of connections.
2325 IPv4 nexthop to fwd packets to.
2327 IPv6 nexthop to fwd packets to.
2332 argument can be used with the following actions:
2333 .Cm nat, pipe , queue, divert, tee, netgraph, ngtee, fwd, skipto, setfib,
2341 action, the user should be aware that the code will walk the ruleset
2342 up to a rule equal to, or past, the given number.
2346 Section for example usage of tables and the tablearg keyword.
2348 Each rule or table belongs to one of 32 different
2351 Set 31 is reserved for the default rule.
2353 By default, rules or tables are put in set 0, unless you use the
2355 attribute when adding a new rule or table.
2356 Sets can be individually and atomically enabled or disabled,
2357 so this mechanism permits an easy way to store multiple configurations
2358 of the firewall and quickly (and atomically) switch between them.
2360 By default, tables from set 0 are referenced when adding rule with
2361 table opcodes regardless of rule set.
2362 This behavior can be changed by setting
2363 .Va net.inet.ip.fw.tables_sets
2365 Rule's set will then be used for table references.
2367 The command to enable/disable sets is
2368 .Bd -ragged -offset indent
2370 .Cm set Oo Cm disable Ar number ... Oc Op Cm enable Ar number ...
2377 sections can be specified.
2378 Command execution is atomic on all the sets specified in the command.
2379 By default, all sets are enabled.
2381 When you disable a set, its rules behave as if they do not exist
2382 in the firewall configuration, with only one exception:
2383 .Bd -ragged -offset indent
2384 dynamic rules created from a rule before it had been disabled
2385 will still be active until they expire.
2387 dynamic rules you have to explicitly delete the parent rule
2388 which generated them.
2391 The set number of rules can be changed with the command
2392 .Bd -ragged -offset indent
2395 .Brq Cm rule Ar rule-number | old-set
2399 Also, you can atomically swap two rulesets with the command
2400 .Bd -ragged -offset indent
2402 .Cm set swap Ar first-set second-set
2407 Section on some possible uses of sets of rules.
2408 .Sh STATEFUL FIREWALL
2409 Stateful operation is a way for the firewall to dynamically
2410 create rules for specific flows when packets that
2411 match a given pattern are detected.
2412 Support for stateful
2413 operation comes through the
2414 .Cm check-state , keep-state , record-state , limit
2420 Dynamic rules are created when a packet matches a
2426 rule, causing the creation of a
2428 rule which will match all and only packets with
2432 .Em src-ip/src-port dst-ip/dst-port
2437 are used here only to denote the initial match addresses, but they
2438 are completely equivalent afterwards).
2444 This name is used in matching together with addresses, ports and protocol.
2445 Dynamic rules will be checked at the first
2446 .Cm check-state, keep-state
2449 occurrence, and the action performed upon a match will be the same
2450 as in the parent rule.
2452 Note that no additional attributes other than protocol and IP addresses
2453 and ports and :flowname are checked on dynamic rules.
2455 The typical use of dynamic rules is to keep a closed firewall configuration,
2456 but let the first TCP SYN packet from the inside network install a
2457 dynamic rule for the flow so that packets belonging to that session
2458 will be allowed through the firewall:
2460 .Dl "ipfw add check-state :OUTBOUND"
2461 .Dl "ipfw add allow tcp from my-subnet to any setup keep-state :OUTBOUND"
2462 .Dl "ipfw add deny tcp from any to any"
2464 A similar approach can be used for UDP, where an UDP packet coming
2465 from the inside will install a dynamic rule to let the response through
2468 .Dl "ipfw add check-state :OUTBOUND"
2469 .Dl "ipfw add allow udp from my-subnet to any keep-state :OUTBOUND"
2470 .Dl "ipfw add deny udp from any to any"
2472 Dynamic rules expire after some time, which depends on the status
2473 of the flow and the setting of some
2477 .Sx SYSCTL VARIABLES
2479 For TCP sessions, dynamic rules can be instructed to periodically
2480 send keepalive packets to refresh the state of the rule when it is
2485 for more examples on how to use dynamic rules.
2486 .Sh TRAFFIC SHAPER (DUMMYNET) CONFIGURATION
2488 is also the user interface for the
2490 traffic shaper, packet scheduler and network emulator, a subsystem that
2491 can artificially queue, delay or drop packets
2492 emulating the behaviour of certain network links
2493 or queueing systems.
2496 operates by first using the firewall to select packets
2497 using any match pattern that can be used in
2500 Matching packets are then passed to either of two
2501 different objects, which implement the traffic regulation:
2502 .Bl -hang -offset XXXX
2508 with given bandwidth and propagation delay,
2509 driven by a FIFO scheduler and a single queue with programmable
2510 queue size and packet loss rate.
2511 Packets are appended to the queue as they come out from
2513 and then transferred in FIFO order to the link at the desired rate.
2517 is an abstraction used to implement packet scheduling
2518 using one of several packet scheduling algorithms.
2521 are first grouped into flows according to a mask on the 5-tuple.
2522 Flows are then passed to the scheduler associated to the
2524 and each flow uses scheduling parameters (weight and others)
2525 as configured in the
2528 A scheduler in turn is connected to an emulated link,
2529 and arbitrates the link's bandwidth among backlogged flows according to
2530 weights and to the features of the scheduling algorithm in use.
2535 can be used to set hard limits to the bandwidth that a flow can use, whereas
2537 can be used to determine how different flows share the available bandwidth.
2539 A graphical representation of the binding of queues,
2540 flows, schedulers and links is below.
2541 .Bd -literal -offset indent
2542 (flow_mask|sched_mask) sched_mask
2543 +---------+ weight Wx +-------------+
2544 | |->-[flow]-->--| |-+
2545 -->--| QUEUE x | ... | | |
2546 | |->-[flow]-->--| SCHEDuler N | |
2548 ... | +--[LINK N]-->--
2549 +---------+ weight Wy | | +--[LINK N]-->--
2550 | |->-[flow]-->--| | |
2551 -->--| QUEUE y | ... | | |
2552 | |->-[flow]-->--| | |
2553 +---------+ +-------------+ |
2556 It is important to understand the role of the SCHED_MASK
2557 and FLOW_MASK, which are configured through the commands
2558 .Dl "ipfw sched N config mask SCHED_MASK ..."
2560 .Dl "ipfw queue X config mask FLOW_MASK ..." .
2562 The SCHED_MASK is used to assign flows to one or more
2563 scheduler instances, one for each
2564 value of the packet's 5-tuple after applying SCHED_MASK.
2565 As an example, using ``src-ip 0xffffff00'' creates one instance
2566 for each /24 destination subnet.
2568 The FLOW_MASK, together with the SCHED_MASK, is used to split
2570 As an example, using
2571 ``src-ip 0x000000ff''
2572 together with the previous SCHED_MASK makes a flow for
2573 each individual source address.
2574 In turn, flows for each /24
2575 subnet will be sent to the same scheduler instance.
2577 The above diagram holds even for the
2579 case, with the only restriction that a
2581 only supports a SCHED_MASK, and forces the use of a FIFO
2582 scheduler (these are for backward compatibility reasons;
2583 in fact, internally, a
2585 pipe is implemented exactly as above).
2587 There are two modes of
2595 mode tries to emulate a real link: the
2597 scheduler ensures that the packet will not leave the pipe faster than it
2598 would on the real link with a given bandwidth.
2601 mode allows certain packets to bypass the
2603 scheduler (if packet flow does not exceed pipe's bandwidth).
2604 This is the reason why the
2606 mode requires less CPU cycles per packet (on average) and packet latency
2607 can be significantly lower in comparison to a real link with the same
2613 mode can be enabled by setting the
2614 .Va net.inet.ip.dummynet.io_fast
2616 variable to a non-zero value.
2618 .Ss PIPE, QUEUE AND SCHEDULER CONFIGURATION
2624 configuration commands are the following:
2625 .Bd -ragged -offset indent
2626 .Cm pipe Ar number Cm config Ar pipe-configuration
2628 .Cm queue Ar number Cm config Ar queue-configuration
2630 .Cm sched Ar number Cm config Ar sched-configuration
2633 The following parameters can be configured for a pipe:
2635 .Bl -tag -width indent -compact
2636 .It Cm bw Ar bandwidth | device
2637 Bandwidth, measured in
2640 .Brq Cm bit/s | Byte/s .
2643 A value of 0 (default) means unlimited bandwidth.
2644 The unit must immediately follow the number, as in
2646 .Dl "ipfw pipe 1 config bw 300Kbit/s"
2648 If a device name is specified instead of a numeric value, as in
2650 .Dl "ipfw pipe 1 config bw tun0"
2652 then the transmit clock is supplied by the specified device.
2653 At the moment only the
2655 device supports this
2656 functionality, for use in conjunction with
2659 .It Cm delay Ar ms-delay
2660 Propagation delay, measured in milliseconds.
2661 The value is rounded to the next multiple of the clock tick
2662 (typically 10ms, but it is a good practice to run kernels
2664 .Dq "options HZ=1000"
2666 the granularity to 1ms or less).
2667 The default value is 0, meaning no delay.
2669 .It Cm burst Ar size
2670 If the data to be sent exceeds the pipe's bandwidth limit
2671 (and the pipe was previously idle), up to
2673 bytes of data are allowed to bypass the
2675 scheduler, and will be sent as fast as the physical link allows.
2676 Any additional data will be transmitted at the rate specified
2680 The burst size depends on how long the pipe has been idle;
2681 the effective burst size is calculated as follows:
2688 .It Cm profile Ar filename
2689 A file specifying the additional overhead incurred in the transmission
2690 of a packet on the link.
2692 Some link types introduce extra delays in the transmission
2693 of a packet, e.g., because of MAC level framing, contention on
2694 the use of the channel, MAC level retransmissions and so on.
2695 From our point of view, the channel is effectively unavailable
2696 for this extra time, which is constant or variable depending
2698 Additionally, packets may be dropped after this
2699 time (e.g., on a wireless link after too many retransmissions).
2700 We can model the additional delay with an empirical curve
2701 that represents its distribution.
2702 .Bd -literal -offset indent
2703 cumulative probability
2713 +-------*------------------->
2716 The empirical curve may have both vertical and horizontal lines.
2717 Vertical lines represent constant delay for a range of
2719 Horizontal lines correspond to a discontinuity in the delay
2720 distribution: the pipe will use the largest delay for a
2723 The file format is the following, with whitespace acting as
2724 a separator and '#' indicating the beginning a comment:
2725 .Bl -tag -width indent
2726 .It Cm name Ar identifier
2727 optional name (listed by "ipfw pipe show")
2728 to identify the delay distribution;
2730 the bandwidth used for the pipe.
2731 If not specified here, it must be present
2732 explicitly as a configuration parameter for the pipe;
2733 .It Cm loss-level Ar L
2734 the probability above which packets are lost.
2735 (0.0 <= L <= 1.0, default 1.0 i.e., no loss);
2737 the number of samples used in the internal
2738 representation of the curve (2..1024; default 100);
2739 .It Cm "delay prob" | "prob delay"
2740 One of these two lines is mandatory and defines
2741 the format of the following lines with data points.
2743 2 or more lines representing points in the curve,
2744 with either delay or probability first, according
2745 to the chosen format.
2746 The unit for delay is milliseconds.
2747 Data points do not need to be sorted.
2748 Also, the number of actual lines can be different
2749 from the value of the "samples" parameter:
2751 utility will sort and interpolate
2752 the curve as needed.
2755 Example of a profile file:
2756 .Bd -literal -offset indent
2761 0 200 # minimum overhead is 200ms
2767 #configuration file end
2771 The following parameters can be configured for a queue:
2773 .Bl -tag -width indent -compact
2774 .It Cm pipe Ar pipe_nr
2775 Connects a queue to the specified pipe.
2776 Multiple queues (with the same or different weights) can be connected to
2777 the same pipe, which specifies the aggregate rate for the set of queues.
2779 .It Cm weight Ar weight
2780 Specifies the weight to be used for flows matching this queue.
2781 The weight must be in the range 1..100, and defaults to 1.
2784 The following case-insensitive parameters can be configured for a
2787 .Bl -tag -width indent -compact
2788 .It Cm type Ar {fifo | wf2q+ | rr | qfq | fq_codel | fq_pie}
2789 specifies the scheduling algorithm to use.
2790 .Bl -tag -width indent -compact
2792 is just a FIFO scheduler (which means that all packets
2793 are stored in the same queue as they arrive to the scheduler).
2794 FIFO has O(1) per-packet time complexity, with very low
2795 constants (estimate 60-80ns on a 2GHz desktop machine)
2796 but gives no service guarantees.
2798 implements the WF2Q+ algorithm, which is a Weighted Fair Queueing
2799 algorithm which permits flows to share bandwidth according to
2801 Note that weights are not priorities; even a flow
2802 with a minuscule weight will never starve.
2803 WF2Q+ has O(log N) per-packet processing cost, where N is the number
2804 of flows, and is the default algorithm used by previous versions
2807 implements the Deficit Round Robin algorithm, which has O(1) processing
2808 costs (roughly, 100-150ns per packet)
2809 and permits bandwidth allocation according to weights, but
2810 with poor service guarantees.
2812 implements the QFQ algorithm, which is a very fast variant of
2813 WF2Q+, with similar service guarantees and O(1) processing
2814 costs (roughly, 200-250ns per packet).
2816 implements the FQ-CoDel (FlowQueue-CoDel) scheduler/AQM algorithm, which
2817 uses a modified Deficit Round Robin scheduler to manage two lists of sub-queues
2818 (old sub-queues and new sub-queues) for providing brief periods of priority to
2819 lightweight or short burst flows.
2820 By default, the total number of sub-queues is 1024.
2821 FQ-CoDel's internal, dynamically
2822 created sub-queues are controlled by separate instances of CoDel AQM.
2824 implements the FQ-PIE (FlowQueue-PIE) scheduler/AQM algorithm, which similar to
2826 but uses per sub-queue PIE AQM instance to control the queue delay.
2830 inherits AQM parameters and options from
2834 inherits AQM parameters and options from
2837 Additionally, both of
2841 have shared scheduler parameters which are:
2842 .Bl -tag -width indent
2845 specifies the quantum (credit) of the scheduler.
2847 is the number of bytes a queue can serve before being moved to the tail
2849 The default is 1514 bytes, and the maximum acceptable value
2853 specifies the hard size limit (in unit of packets) of all queues managed by an
2854 instance of the scheduler.
2855 The default value of
2857 is 10240 packets, and the maximum acceptable value is 20480 packets.
2860 specifies the total number of flow queues (sub-queues) that fq_*
2861 creates and manages.
2862 By default, 1024 sub-queues are created when an instance
2863 of the fq_{codel/pie} scheduler is created.
2864 The maximum acceptable value is
2868 Note that any token after
2872 is considered a parameter for fq_{codel/pie}.
2873 So, ensure all scheduler
2874 configuration options not related to fq_{codel/pie} are written before
2879 In addition to the type, all parameters allowed for a pipe can also
2880 be specified for a scheduler.
2882 Finally, the following parameters can be configured for both
2885 .Bl -tag -width XXXX -compact
2886 .It Cm buckets Ar hash-table-size
2887 Specifies the size of the hash table used for storing the
2889 Default value is 64 controlled by the
2892 .Va net.inet.ip.dummynet.hash_size ,
2893 allowed range is 16 to 65536.
2895 .It Cm mask Ar mask-specifier
2896 Packets sent to a given pipe or queue by an
2898 rule can be further classified into multiple flows, each of which is then
2902 A flow identifier is constructed by masking the IP addresses,
2903 ports and protocol types as specified with the
2905 options in the configuration of the pipe or queue.
2906 For each different flow identifier, a new pipe or queue is created
2907 with the same parameters as the original object, and matching packets
2912 are used, each flow will get the same bandwidth as defined by the pipe,
2915 are used, each flow will share the parent's pipe bandwidth evenly
2916 with other flows generated by the same queue (note that other queues
2917 with different weights might be connected to the same pipe).
2919 Available mask specifiers are a combination of one or more of the following:
2921 .Cm dst-ip Ar mask ,
2922 .Cm dst-ip6 Ar mask ,
2923 .Cm src-ip Ar mask ,
2924 .Cm src-ip6 Ar mask ,
2925 .Cm dst-port Ar mask ,
2926 .Cm src-port Ar mask ,
2927 .Cm flow-id Ar mask ,
2932 where the latter means all bits in all fields are significant.
2935 When a packet is dropped by a
2937 queue or pipe, the error
2938 is normally reported to the caller routine in the kernel, in the
2939 same way as it happens when a device queue fills up.
2941 option reports the packet as successfully delivered, which can be
2942 needed for some experimental setups where you want to simulate
2943 loss or congestion at a remote router.
2945 .It Cm plr Ar packet-loss-rate
2948 .Ar packet-loss-rate
2949 is a floating-point number between 0 and 1, with 0 meaning no
2950 loss, 1 meaning 100% loss.
2951 The loss rate is internally represented on 31 bits.
2953 .It Cm queue Brq Ar slots | size Ns Cm Kbytes
2958 Default value is 50 slots, which
2959 is the typical queue size for Ethernet devices.
2960 Note that for slow speed links you should keep the queue
2961 size short or your traffic might be affected by a significant
2963 E.g., 50 max-sized Ethernet packets (1500 bytes) mean 600Kbit
2964 or 20s of queue on a 30Kbit/s pipe.
2965 Even worse effects can result if you get packets from an
2966 interface with a much larger MTU, e.g.\& the loopback interface
2967 with its 16KB packets.
2971 .Em net.inet.ip.dummynet.pipe_byte_limit
2973 .Em net.inet.ip.dummynet.pipe_slot_limit
2974 control the maximum lengths that can be specified.
2976 .It Cm red | gred Ar w_q Ns / Ns Ar min_th Ns / Ns Ar max_th Ns / Ns Ar max_p
2978 Make use of the RED (Random Early Detection) queue management algorithm.
2983 point numbers between 0 and 1 (inclusive), while
2987 are integer numbers specifying thresholds for queue management
2988 (thresholds are computed in bytes if the queue has been defined
2989 in bytes, in slots otherwise).
2990 The two parameters can also be of the same value if needed.
2993 also supports the gentle RED variant (gred) and ECN (Explicit Congestion
2994 Notification) as optional.
2997 variables can be used to control the RED behaviour:
2998 .Bl -tag -width indent
2999 .It Va net.inet.ip.dummynet.red_lookup_depth
3000 specifies the accuracy in computing the average queue
3001 when the link is idle (defaults to 256, must be greater than zero)
3002 .It Va net.inet.ip.dummynet.red_avg_pkt_size
3003 specifies the expected average packet size (defaults to 512, must be
3005 .It Va net.inet.ip.dummynet.red_max_pkt_size
3006 specifies the expected maximum packet size, only used when queue
3007 thresholds are in bytes (defaults to 1500, must be greater than zero).
3010 .It Cm codel Oo Cm target Ar time Oc Oo Cm interval Ar time Oc Oo Cm ecn |
3012 Make use of the CoDel (Controlled-Delay) queue management algorithm.
3014 is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
3015 microseconds (us) can be specified instead.
3016 CoDel drops or marks (ECN) packets
3017 depending on packet sojourn time in the queue.
3020 (5ms by default) is the minimum acceptable persistent queue delay that CoDel
3022 CoDel does not drop packets directly after packets sojourn time becomes
3029 (100ms default) before dropping.
3032 should be set to maximum RTT for all expected connections.
3034 enables (disabled by default) packet marking (instead of dropping) for
3035 ECN-enabled TCP flows when queue delay becomes high.
3037 Note that any token after
3039 is considered a parameter for CoDel.
3040 So, ensure all pipe/queue
3041 configuration options are written before
3048 .Va net.inet.ip.dummynet.codel.target
3050 .Va net.inet.ip.dummynet.codel.interval
3051 can be used to set CoDel default parameters.
3053 .It Cm pie Oo Cm target Ar time Oc Oo Cm tupdate Ar time Oc Oo
3054 .Cm alpha Ar n Oc Oo Cm beta Ar n Oc Oo Cm max_burst Ar time Oc Oo
3055 .Cm max_ecnth Ar n Oc Oo Cm ecn | Cm noecn Oc Oo Cm capdrop |
3056 .Cm nocapdrop Oc Oo Cm drand | Cm nodrand Oc Oo Cm onoff
3057 .Oc Oo Cm dre | Cm ts Oc
3058 Make use of the PIE (Proportional Integral controller Enhanced) queue management
3060 PIE drops or marks packets depending on a calculated drop probability during
3061 en-queue process, with the aim of achieving high throughput while keeping queue
3063 At regular time intervals of
3066 (15ms by default) a background process (re)calculates the probability based on queue delay
3070 (15ms by default) and queue delay trends.
3071 PIE approximates current queue
3072 delay by using a departure rate estimation method, or (optionally) by using a
3073 packet timestamp method similar to CoDel.
3075 is interpreted as milliseconds by default but seconds (s), milliseconds (ms) or
3076 microseconds (us) can be specified instead.
3077 The other PIE parameters and options are as follows:
3078 .Bl -tag -width indent
3081 is a floating point number between 0 and 7 which specifies the weight of queue
3082 delay deviations that is used in drop probability calculation.
3083 0.125 is the default.
3086 is a floating point number between 0 and 7 which specifies is the weight of queue
3087 delay trend that is used in drop probability calculation.
3088 1.25 is the default.
3089 .It Cm max_burst Ar time
3090 The maximum period of time that PIE does not drop/mark packets.
3092 default and 10s is the maximum value.
3093 .It Cm max_ecnth Ar n
3094 Even when ECN is enabled, PIE drops packets instead of marking them when drop
3095 probability becomes higher than ECN probability threshold
3097 , the default is 0.1 (i.e 10%) and 1 is the maximum value.
3099 enable or disable ECN marking for ECN-enabled TCP flows.
3100 Disabled by default.
3101 .It Cm capdrop | nocapdrop
3102 enable or disable cap drop adjustment.
3103 Cap drop adjustment is enabled by default.
3104 .It Cm drand | nodrand
3105 enable or disable drop probability de-randomisation.
3106 De-randomisation eliminates
3107 the problem of dropping packets too close or too far.
3108 De-randomisation is enabled by default.
3110 enable turning PIE on and off depending on queue load.
3111 If this option is enabled,
3112 PIE turns on when over 1/3 of queue becomes full.
3113 This option is disabled by
3116 Calculate queue delay using departure rate estimation
3124 Note that any token after
3126 is considered a parameter for PIE.
3127 So ensure all pipe/queue
3128 the configuration options are written before
3132 variables can be used to control the
3136 .Sx SYSCTL VARIABLES
3137 section for more details.
3140 When used with IPv6 data,
3142 currently has several limitations.
3143 Information necessary to route link-local packets to an
3144 interface is not available after processing by
3146 so those packets are dropped in the output path.
3147 Care should be taken to ensure that link-local packets are not passed to
3150 Here are some important points to consider when designing your
3154 Remember that you filter both packets going
3158 Most connections need packets going in both directions.
3160 Remember to test very carefully.
3161 It is a good idea to be near the console when doing this.
3162 If you cannot be near the console,
3163 use an auto-recovery script such as the one in
3164 .Pa /usr/share/examples/ipfw/change_rules.sh .
3166 Do not forget the loopback interface.
3171 There are circumstances where fragmented datagrams are unconditionally
3173 TCP packets are dropped if they do not contain at least 20 bytes of
3174 TCP header, UDP packets are dropped if they do not contain a full 8
3175 byte UDP header, and ICMP packets are dropped if they do not contain
3176 4 bytes of ICMP header, enough to specify the ICMP type, code, and
3178 These packets are simply logged as
3180 since there may not be enough good data in the packet to produce a
3181 meaningful log entry.
3183 Another type of packet is unconditionally dropped, a TCP packet with a
3184 fragment offset of one.
3185 This is a valid packet, but it only has one use, to try
3186 to circumvent firewalls.
3187 When logging is enabled, these packets are
3188 reported as being dropped by rule -1.
3190 If you are logged in over a network, loading the
3194 is probably not as straightforward as you would think.
3195 The following command line is recommended:
3196 .Bd -literal -offset indent
3198 ipfw add 32000 allow ip from any to any
3201 Along the same lines, doing an
3202 .Bd -literal -offset indent
3206 in similar surroundings is also a bad idea.
3210 filter list may not be modified if the system security level
3211 is set to 3 or higher
3214 for information on system security levels).
3216 .Sh PACKET DIVERSION
3219 socket bound to the specified port will receive all packets
3220 diverted to that port.
3221 If no socket is bound to the destination port, or if the divert module is
3222 not loaded, or if the kernel was not compiled with divert socket support,
3223 the packets are dropped.
3224 .Sh NETWORK ADDRESS TRANSLATION (NAT)
3226 support in-kernel NAT using the kernel version of
3230 should be loaded or kernel should have
3231 .Cm options IPFIREWALL_NAT
3234 The nat configuration command is the following:
3235 .Bd -ragged -offset indent
3240 .Ar nat-configuration
3244 The following parameters can be configured:
3245 .Bl -tag -width indent
3246 .It Cm ip Ar ip_address
3247 Define an ip address to use for aliasing.
3249 Use ip address of NIC for aliasing, dynamically changing
3250 it if NIC's ip address changes.
3252 Enable logging on this nat instance.
3254 Deny any incoming connection from outside world.
3256 Try to leave the alias port numbers unchanged from
3257 the actual local port numbers.
3259 Traffic on the local network not originating from a RFC 1918
3260 unregistered address spaces will be ignored.
3262 Like unreg_only, but includes the RFC 6598 (Carrier Grade NAT)
3265 Reset table of the packet aliasing engine on address change.
3267 Reverse the way libalias handles aliasing.
3269 Obey transparent proxy rules only, packet aliasing is not performed.
3271 Skip instance in case of global state lookup (see below).
3274 Some specials value can be supplied instead of
3276 .Bl -tag -width indent
3278 Looks up translation state in all configured nat instances.
3279 If an entry is found, packet is aliased according to that entry.
3280 If no entry was found in any of the instances, packet is passed unchanged,
3281 and no new entry will be created.
3283 .Sx MULTIPLE INSTANCES
3286 for more information.
3288 Uses argument supplied in lookup table.
3291 section below for more information on lookup tables.
3294 To let the packet continue after being (de)aliased, set the sysctl variable
3295 .Va net.inet.ip.fw.one_pass
3297 For more information about aliasing modes, refer to
3301 for some examples about nat usage.
3302 .Ss REDIRECT AND LSNAT SUPPORT IN IPFW
3303 Redirect and LSNAT support follow closely the syntax used in
3307 for some examples on how to do redirect and lsnat.
3308 .Ss SCTP NAT SUPPORT
3309 SCTP nat can be configured in a similar manner to TCP through the
3312 The main difference is that
3314 does not do port translation.
3315 Since the local and global side ports will be the same,
3316 there is no need to specify both.
3317 Ports are redirected as follows:
3318 .Bd -ragged -offset indent
3324 .Cm redirect_port sctp
3325 .Ar ip_address [,addr_list] {[port | port-port] [,ports]}
3331 configuration can be done in real-time through the
3334 All may be changed dynamically, though the hash_table size will only
3339 .Sx SYSCTL VARIABLES
3341 .Sh IPv6/IPv4 NETWORK ADDRESS AND PROTOCOL TRANSLATION
3342 .Ss Stateful translation
3344 supports in-kernel IPv6/IPv4 network address and protocol translation.
3345 Stateful NAT64 translation allows IPv6-only clients to contact IPv4 servers
3346 using unicast TCP, UDP or ICMP protocols.
3347 One or more IPv4 addresses assigned to a stateful NAT64 translator are shared
3348 among several IPv6-only clients.
3349 When stateful NAT64 is used in conjunction with DNS64, no changes are usually
3350 required in the IPv6 client or the IPv4 server.
3353 should be loaded or kernel should have
3354 .Cm options IPFIREWALL_NAT64
3355 to be able use stateful NAT64 translator.
3357 Stateful NAT64 uses a bunch of memory for several types of objects.
3358 When IPv6 client initiates connection, NAT64 translator creates a host entry
3359 in the states table.
3360 Each host entry uses preallocated IPv4 alias entry.
3361 Each alias entry has a number of ports group entries allocated on demand.
3362 Ports group entries contains connection state entries.
3363 There are several options to control limits and lifetime for these objects.
3365 NAT64 translator follows RFC7915 when does ICMPv6/ICMP translation,
3366 unsupported message types will be silently dropped.
3367 IPv6 needs several ICMPv6 message types to be explicitly allowed for correct
3369 Make sure that ND6 neighbor solicitation (ICMPv6 type 135) and neighbor
3370 advertisement (ICMPv6 type 136) messages will not be handled by translation
3373 After translation NAT64 translator by default sends packets through
3374 corresponding netisr queue.
3375 Thus translator host should be configured as IPv4 and IPv6 router.
3376 Also this means, that a packet is handled by firewall twice.
3377 First time an original packet is handled and consumed by translator,
3378 and then it is handled again as translated packet.
3379 This behavior can be changed by sysctl variable
3380 .Va net.inet.ip.fw.nat64_direct_output .
3381 Also translated packet can be tagged using
3383 rule action, and then matched by
3385 opcode to avoid loops and extra overhead.
3387 The stateful NAT64 configuration command is the following:
3388 .Bd -ragged -offset indent
3397 The following parameters can be configured:
3398 .Bl -tag -width indent
3399 .It Cm prefix4 Ar ipv4_prefix/plen
3400 The IPv4 prefix with mask defines the pool of IPv4 addresses used as
3401 source address after translation.
3402 Stateful NAT64 module translates IPv6 source address of client to one
3403 IPv4 address from this pool.
3404 Note that incoming IPv4 packets that don't have corresponding state entry
3405 in the states table will be dropped by translator.
3406 Make sure that translation rules handle packets, destined to configured prefix.
3407 .It Cm prefix6 Ar ipv6_prefix/length
3408 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3409 to represent IPv4 addresses.
3410 This IPv6 prefix should be configured in DNS64.
3411 The translator implementation follows RFC6052, that restricts the length of
3412 prefixes to one of following: 32, 40, 48, 56, 64, or 96.
3413 The Well-Known IPv6 Prefix 64:ff9b:: must be 96 bits long.
3416 prefix can be used to handle several IPv6 prefixes with one NAT64 instance.
3417 The NAT64 instance will determine a destination IPv4 address from prefix
3419 .It Cm states_chunks Ar number
3420 The number of states chunks in single ports group.
3421 Each ports group by default can keep 64 state entries in single chunk.
3422 The above value affects the maximum number of states that can be associated with single IPv4 alias address and port.
3423 The value must be power of 2, and up to 128.
3424 .It Cm host_del_age Ar seconds
3425 The number of seconds until the host entry for a IPv6 client will be deleted
3426 and all its resources will be released due to inactivity.
3429 .It Cm pg_del_age Ar seconds
3430 The number of seconds until a ports group with unused state entries will
3434 .It Cm tcp_syn_age Ar seconds
3435 The number of seconds while a state entry for TCP connection with only SYN
3437 If TCP connection establishing will not be finished,
3438 state entry will be deleted.
3441 .It Cm tcp_est_age Ar seconds
3442 The number of seconds while a state entry for established TCP connection
3446 .It Cm tcp_close_age Ar seconds
3447 The number of seconds while a state entry for closed TCP connection
3449 Keeping state entries for closed connections is needed, because IPv4 servers
3450 typically keep closed connections in a TIME_WAIT state for a several minutes.
3451 Since translator's IPv4 addresses are shared among all IPv6 clients,
3452 new connections from the same addresses and ports may be rejected by server,
3453 because these connections are still in a TIME_WAIT state.
3454 Keeping them in translator's state table protects from such rejects.
3457 .It Cm udp_age Ar seconds
3458 The number of seconds while translator keeps state entry in a waiting for
3459 reply to the sent UDP datagram.
3462 .It Cm icmp_age Ar seconds
3463 The number of seconds while translator keeps state entry in a waiting for
3464 reply to the sent ICMP message.
3468 Turn on logging of all handled packets via BPF through
3472 is a pseudo interface and can be created after a boot manually with
3475 Note that it has different purpose than
3478 Translators sends to BPF an additional information with each packet.
3481 you are able to see each handled packet before and after translation.
3483 Turn off logging of all handled packets via BPF.
3484 .It Cm allow_private
3485 Turn on processing private IPv4 addresses.
3486 By default IPv6 packets with destinations mapped to private address ranges
3487 defined by RFC1918 are not processed.
3488 .It Cm -allow_private
3489 Turn off private address handling in
3494 To inspect a states table of stateful NAT64 the following command can be used:
3495 .Bd -ragged -offset indent
3503 Stateless NAT64 translator doesn't use a states table for translation
3504 and converts IPv4 addresses to IPv6 and vice versa solely based on the
3505 mappings taken from configured lookup tables.
3506 Since a states table doesn't used by stateless translator,
3507 it can be configured to pass IPv4 clients to IPv6-only servers.
3509 The stateless NAT64 configuration command is the following:
3510 .Bd -ragged -offset indent
3519 The following parameters can be configured:
3520 .Bl -tag -width indent
3521 .It Cm prefix6 Ar ipv6_prefix/length
3522 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3523 to represent IPv4 addresses.
3524 This IPv6 prefix should be configured in DNS64.
3525 .It Cm table4 Ar table46
3528 contains mapping how IPv4 addresses should be translated to IPv6 addresses.
3529 .It Cm table6 Ar table64
3532 contains mapping how IPv6 addresses should be translated to IPv4 addresses.
3534 Turn on logging of all handled packets via BPF through
3538 Turn off logging of all handled packets via BPF.
3539 .It Cm allow_private
3540 Turn on processing private IPv4 addresses.
3541 By default IPv6 packets with destinations mapped to private address ranges
3542 defined by RFC1918 are not processed.
3543 .It Cm -allow_private
3544 Turn off private address handling in
3549 Note that the behavior of stateless translator with respect to not matched
3550 packets differs from stateful translator.
3551 If corresponding addresses was not found in the lookup tables, the packet
3552 will not be dropped and the search continues.
3554 .Ss XLAT464 CLAT translation
3555 XLAT464 CLAT NAT64 translator implements client-side stateless translation as
3556 defined in RFC6877 and is very similar to statless NAT64 translator
3558 Instead of lookup tables it uses one-to-one mapping between IPv4 and IPv6
3559 addresses using configured prefixes.
3560 This mode can be used as a replacement of DNS64 service for applications
3561 that are not using it (e.g. VoIP) allowing them to access IPv4-only Internet
3562 over IPv6-only networks with help of remote NAT64 translator.
3564 The CLAT NAT64 configuration command is the following:
3565 .Bd -ragged -offset indent
3574 The following parameters can be configured:
3575 .Bl -tag -width indent
3576 .It Cm clat_prefix Ar ipv6_prefix/length
3577 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3578 to represent source IPv4 addresses.
3579 .It Cm plat_prefix Ar ipv6_prefix/length
3580 The IPv6 prefix defines IPv4-embedded IPv6 addresses used by translator
3581 to represent destination IPv4 addresses.
3582 This IPv6 prefix should be configured on a remote NAT64 translator.
3584 Turn on logging of all handled packets via BPF through
3588 Turn off logging of all handled packets via BPF.
3589 .It Cm allow_private
3590 Turn on processing private IPv4 addresses.
3593 instance will not process IPv4 packets with destination address from private
3594 ranges as defined in RFC1918.
3595 .It Cm -allow_private
3596 Turn off private address handling in
3601 Note that the behavior of CLAT translator with respect to not matched
3602 packets differs from stateful translator.
3603 If corresponding addresses were not matched against prefixes configured,
3604 the packet will not be dropped and the search continues.
3605 .Sh IPv6-to-IPv6 NETWORK PREFIX TRANSLATION (NPTv6)
3607 supports in-kernel IPv6-to-IPv6 network prefix translation as described
3611 should be loaded or kernel should has
3612 .Cm options IPFIREWALL_NPTV6
3613 to be able use NPTv6 translator.
3615 The NPTv6 configuration command is the following:
3616 .Bd -ragged -offset indent
3625 The following parameters can be configured:
3626 .Bl -tag -width indent
3627 .It Cm int_prefix Ar ipv6_prefix
3628 IPv6 prefix used in internal network.
3629 NPTv6 module translates source address when it matches this prefix.
3630 .It Cm ext_prefix Ar ipv6_prefix
3631 IPv6 prefix used in external network.
3632 NPTv6 module translates destination address when it matches this prefix.
3633 .It Cm ext_if Ar nic
3634 The NPTv6 module will use first global IPv6 address from interface
3637 It can be useful when IPv6 prefix of external network is dynamically obtained.
3641 options are mutually exclusive.
3642 .It Cm prefixlen Ar length
3643 The length of specified IPv6 prefixes.
3644 It must be in range from 8 to 64.
3647 Note that the prefix translation rules are silently ignored when IPv6 packet
3648 forwarding is disabled.
3649 To enable the packet forwarding, set the sysctl variable
3650 .Va net.inet6.ip6.forwarding
3653 To let the packet continue after being translated, set the sysctl variable
3654 .Va net.inet.ip.fw.one_pass
3657 Tunables can be set in
3663 before ipfw module gets loaded.
3664 .Bl -tag -width indent
3665 .It Va net.inet.ip.fw.default_to_accept: No 0
3666 Defines ipfw last rule behavior.
3667 This value overrides
3668 .Cd "options IPFW_DEFAULT_TO_(ACCEPT|DENY)"
3669 from kernel configuration file.
3670 .It Va net.inet.ip.fw.tables_max: No 128
3671 Defines number of tables available in ipfw.
3672 Number cannot exceed 65534.
3674 .Sh SYSCTL VARIABLES
3677 variables controls the behaviour of the firewall and
3679 .Pq Nm dummynet , bridge , sctp nat .
3680 These are shown below together with their default value
3681 (but always check with the
3683 command what value is actually in use) and meaning:
3684 .Bl -tag -width indent
3685 .It Va net.inet.ip.alias.sctp.accept_global_ootb_addip: No 0
3688 responds to receipt of global OOTB ASCONF-AddIP:
3689 .Bl -tag -width indent
3691 No response (unless a partially matching association exists -
3692 ports and vtags match but global address does not)
3695 will accept and process all OOTB global AddIP messages.
3698 Option 1 should never be selected as this forms a security risk.
3700 establish multiple fake associations by sending AddIP messages.
3701 .It Va net.inet.ip.alias.sctp.chunk_proc_limit: No 5
3702 Defines the maximum number of chunks in an SCTP packet that will be
3704 packet that matches an existing association.
3705 This value is enforced to be greater or equal than
3706 .Cm net.inet.ip.alias.sctp.initialising_chunk_proc_limit .
3708 a DoS risk yet setting too low a value may result in
3709 important control chunks in
3710 the packet not being located and parsed.
3711 .It Va net.inet.ip.alias.sctp.error_on_ootb: No 1
3714 responds to any Out-of-the-Blue (OOTB) packets with ErrorM packets.
3715 An OOTB packet is a packet that arrives with no existing association
3718 and is not an INIT or ASCONF-AddIP packet:
3719 .Bl -tag -width indent
3721 ErrorM is never sent in response to OOTB packets.
3723 ErrorM is only sent to OOTB packets received on the local side.
3725 ErrorM is sent to the local side and on the global side ONLY if there is a
3726 partial match (ports and vtags match but the source global IP does not).
3727 This value is only useful if the
3729 is tracking global IP addresses.
3731 ErrorM is sent in response to all OOTB packets on both
3732 the local and global side
3736 At the moment the default is 0, since the ErrorM packet is not yet
3737 supported by most SCTP stacks.
3738 When it is supported, and if not tracking
3739 global addresses, we recommend setting this value to 1 to allow
3740 multi-homed local hosts to function with the
3742 To track global addresses, we recommend setting this value to 2 to
3743 allow global hosts to be informed when they need to (re)send an
3745 Value 3 should never be chosen (except for debugging) as the
3747 will respond to all OOTB global packets (a DoS risk).
3748 .It Va net.inet.ip.alias.sctp.hashtable_size: No 2003
3749 Size of hash tables used for
3751 lookups (100 < prime_number > 1000001).
3754 size for any future created
3756 instance and therefore must be set prior to creating a
3759 The table sizes may be changed to suit specific needs.
3760 If there will be few
3761 concurrent associations, and memory is scarce, you may make these smaller.
3762 If there will be many thousands (or millions) of concurrent associations, you
3763 should make these larger.
3764 A prime number is best for the table size.
3766 update function will adjust your input value to the next highest prime number.
3767 .It Va net.inet.ip.alias.sctp.holddown_time: No 0
3768 Hold association in table for this many seconds after receiving a
3770 This allows endpoints to correct shutdown gracefully if a
3771 shutdown_complete is lost and retransmissions are required.
3772 .It Va net.inet.ip.alias.sctp.init_timer: No 15
3773 Timeout value while waiting for (INIT-ACK|AddIP-ACK).
3774 This value cannot be 0.
3775 .It Va net.inet.ip.alias.sctp.initialising_chunk_proc_limit: No 2
3776 Defines the maximum number of chunks in an SCTP packet that will be parsed when
3777 no existing association exists that matches that packet.
3779 will only be an INIT or ASCONF-AddIP packet.
3780 A higher value may become a DoS
3781 risk as malformed packets can consume processing resources.
3782 .It Va net.inet.ip.alias.sctp.param_proc_limit: No 25
3783 Defines the maximum number of parameters within a chunk that will be
3786 As for other similar sysctl variables, larger values pose a DoS risk.
3787 .It Va net.inet.ip.alias.sctp.log_level: No 0
3788 Level of detail in the system log messages (0 \- minimal, 1 \- event,
3789 2 \- info, 3 \- detail, 4 \- debug, 5 \- max debug).
3791 option in high loss environments.
3792 .It Va net.inet.ip.alias.sctp.shutdown_time: No 15
3793 Timeout value while waiting for SHUTDOWN-COMPLETE.
3794 This value cannot be 0.
3795 .It Va net.inet.ip.alias.sctp.track_global_addresses: No 0
3796 Enables/disables global IP address tracking within the
3799 upper limit on the number of addresses tracked for each association:
3800 .Bl -tag -width indent
3802 Global tracking is disabled
3804 Enables tracking, the maximum number of addresses tracked for each
3805 association is limited to this value
3808 This variable is fully dynamic, the new value will be adopted for all newly
3809 arriving associations, existing associations are treated
3810 as they were previously.
3811 Global tracking will decrease the number of collisions within the
3814 of increased processing load, memory usage, complexity, and possible
3817 problems in complex networks with multiple
3819 We recommend not tracking
3820 global IP addresses, this will still result in a fully functional
3822 .It Va net.inet.ip.alias.sctp.up_timer: No 300
3823 Timeout value to keep an association up with no traffic.
3824 This value cannot be 0.
3825 .It Va net.inet.ip.dummynet.codel.interval : No 100000
3828 AQM interval in microseconds.
3829 The value must be in the range 1..5000000.
3830 .It Va net.inet.ip.dummynet.codel.target : No 5000
3833 AQM target delay time in microseconds (the minimum acceptable persistent queue
3835 The value must be in the range 1..5000000.
3836 .It Va net.inet.ip.dummynet.expire : No 1
3837 Lazily delete dynamic pipes/queue once they have no pending traffic.
3838 You can disable this by setting the variable to 0, in which case
3839 the pipes/queues will only be deleted when the threshold is reached.
3840 .It Va net.inet.ip.dummynet.fqcodel.flows : No 1024
3841 Defines the default total number of flow queues (sub-queues) that
3843 creates and manages.
3844 The value must be in the range 1..65536.
3845 .It Va net.inet.ip.dummynet.fqcodel.interval : No 100000
3848 scheduler/AQM interval in microseconds.
3849 The value must be in the range 1..5000000.
3850 .It Va net.inet.ip.dummynet.fqcodel.limit : No 10240
3851 The default hard size limit (in unit of packet) of all queues managed by an
3855 The value must be in the range 1..20480.
3856 .It Va net.inet.ip.dummynet.fqcodel.quantum : No 1514
3857 The default quantum (credit) of the
3860 The value must be in the range 1..9000.
3861 .It Va net.inet.ip.dummynet.fqcodel.target : No 5000
3864 scheduler/AQM target delay time in microseconds (the minimum acceptable
3865 persistent queue delay).
3866 The value must be in the range 1..5000000.
3867 .It Va net.inet.ip.dummynet.fqpie.alpha : No 125
3870 parameter (scaled by 1000) for
3873 The value must be in the range 1..7000.
3874 .It Va net.inet.ip.dummynet.fqpie.beta : No 1250
3877 parameter (scaled by 1000) for
3880 The value must be in the range 1..7000.
3881 .It Va net.inet.ip.dummynet.fqpie.flows : No 1024
3882 Defines the default total number of flow queues (sub-queues) that
3884 creates and manages.
3885 The value must be in the range 1..65536.
3886 .It Va net.inet.ip.dummynet.fqpie.limit : No 10240
3887 The default hard size limit (in unit of packet) of all queues managed by an
3891 The value must be in the range 1..20480.
3892 .It Va net.inet.ip.dummynet.fqpie.max_burst : No 150000
3893 The default maximum period of microseconds that
3895 scheduler/AQM does not drop/mark packets.
3896 The value must be in the range 1..10000000.
3897 .It Va net.inet.ip.dummynet.fqpie.max_ecnth : No 99
3898 The default maximum ECN probability threshold (scaled by 1000) for
3901 The value must be in the range 1..7000.
3902 .It Va net.inet.ip.dummynet.fqpie.quantum : No 1514
3903 The default quantum (credit) of the
3906 The value must be in the range 1..9000.
3907 .It Va net.inet.ip.dummynet.fqpie.target : No 15000
3912 in unit of microsecond.
3913 The value must be in the range 1..5000000.
3914 .It Va net.inet.ip.dummynet.fqpie.tupdate : No 15000
3919 in unit of microsecond.
3920 The value must be in the range 1..5000000.
3921 .It Va net.inet.ip.dummynet.hash_size : No 64
3922 Default size of the hash table used for dynamic pipes/queues.
3923 This value is used when no
3925 option is specified when configuring a pipe/queue.
3926 .It Va net.inet.ip.dummynet.io_fast : No 0
3927 If set to a non-zero value,
3932 operation (see above) is enabled.
3933 .It Va net.inet.ip.dummynet.io_pkt
3934 Number of packets passed to
3936 .It Va net.inet.ip.dummynet.io_pkt_drop
3937 Number of packets dropped by
3939 .It Va net.inet.ip.dummynet.io_pkt_fast
3940 Number of packets bypassed by the
3943 .It Va net.inet.ip.dummynet.max_chain_len : No 16
3944 Target value for the maximum number of pipes/queues in a hash bucket.
3946 .Cm max_chain_len*hash_size
3947 is used to determine the threshold over which empty pipes/queues
3948 will be expired even when
3949 .Cm net.inet.ip.dummynet.expire=0 .
3950 .It Va net.inet.ip.dummynet.red_lookup_depth : No 256
3951 .It Va net.inet.ip.dummynet.red_avg_pkt_size : No 512
3952 .It Va net.inet.ip.dummynet.red_max_pkt_size : No 1500
3953 Parameters used in the computations of the drop probability
3954 for the RED algorithm.
3955 .It Va net.inet.ip.dummynet.pie.alpha : No 125
3958 parameter (scaled by 1000) for
3961 The value must be in the range 1..7000.
3962 .It Va net.inet.ip.dummynet.pie.beta : No 1250
3965 parameter (scaled by 1000) for
3968 The value must be in the range 1..7000.
3969 .It Va net.inet.ip.dummynet.pie.max_burst : No 150000
3970 The default maximum period of microseconds that
3972 AQM does not drop/mark packets.
3973 The value must be in the range 1..10000000.
3974 .It Va net.inet.ip.dummynet.pie.max_ecnth : No 99
3975 The default maximum ECN probability threshold (scaled by 1000) for
3978 The value must be in the range 1..7000.
3979 .It Va net.inet.ip.dummynet.pie.target : No 15000
3984 AQM in unit of microsecond.
3985 The value must be in the range 1..5000000.
3986 .It Va net.inet.ip.dummynet.pie.tupdate : No 15000
3991 AQM in unit of microsecond.
3992 The value must be in the range 1..5000000.
3993 .It Va net.inet.ip.dummynet.pipe_byte_limit : No 1048576
3994 .It Va net.inet.ip.dummynet.pipe_slot_limit : No 100
3995 The maximum queue size that can be specified in bytes or packets.
3996 These limits prevent accidental exhaustion of resources such as mbufs.
3997 If you raise these limits,
3998 you should make sure the system is configured so that sufficient resources
4000 .It Va net.inet.ip.fw.autoinc_step : No 100
4001 Delta between rule numbers when auto-generating them.
4002 The value must be in the range 1..1000.
4003 .It Va net.inet.ip.fw.curr_dyn_buckets : Va net.inet.ip.fw.dyn_buckets
4004 The current number of buckets in the hash table for dynamic rules
4006 .It Va net.inet.ip.fw.debug : No 1
4007 Controls debugging messages produced by
4009 .It Va net.inet.ip.fw.default_rule : No 65535
4010 The default rule number (read-only).
4012 .Nm , the default rule is the last one, so its number
4013 can also serve as the highest number allowed for a rule.
4014 .It Va net.inet.ip.fw.dyn_buckets : No 256
4015 The number of buckets in the hash table for dynamic rules.
4016 Must be a power of 2, up to 65536.
4017 It only takes effect when all dynamic rules have expired, so you
4018 are advised to use a
4020 command to make sure that the hash table is resized.
4021 .It Va net.inet.ip.fw.dyn_count : No 3
4022 Current number of dynamic rules
4024 .It Va net.inet.ip.fw.dyn_keepalive : No 1
4025 Enables generation of keepalive packets for
4027 rules on TCP sessions.
4028 A keepalive is generated to both
4029 sides of the connection every 5 seconds for the last 20
4030 seconds of the lifetime of the rule.
4031 .It Va net.inet.ip.fw.dyn_max : No 8192
4032 Maximum number of dynamic rules.
4033 When you hit this limit, no more dynamic rules can be
4034 installed until old ones expire.
4035 .It Va net.inet.ip.fw.dyn_ack_lifetime : No 300
4036 .It Va net.inet.ip.fw.dyn_syn_lifetime : No 20
4037 .It Va net.inet.ip.fw.dyn_fin_lifetime : No 1
4038 .It Va net.inet.ip.fw.dyn_rst_lifetime : No 1
4039 .It Va net.inet.ip.fw.dyn_udp_lifetime : No 5
4040 .It Va net.inet.ip.fw.dyn_short_lifetime : No 30
4041 These variables control the lifetime, in seconds, of dynamic
4043 Upon the initial SYN exchange the lifetime is kept short,
4044 then increased after both SYN have been seen, then decreased
4045 again during the final FIN exchange or when a RST is received.
4047 .Em dyn_fin_lifetime
4049 .Em dyn_rst_lifetime
4050 must be strictly lower than 5 seconds, the period of
4051 repetition of keepalives.
4052 The firewall enforces that.
4053 .It Va net.inet.ip.fw.dyn_keep_states: No 0
4054 Keep dynamic states on rule/set deletion.
4055 States are relinked to default rule (65535).
4056 This can be handly for ruleset reload.
4057 Turned off by default.
4058 .It Va net.inet.ip.fw.enable : No 1
4059 Enables the firewall.
4060 Setting this variable to 0 lets you run your machine without
4061 firewall even if compiled in.
4062 .It Va net.inet6.ip6.fw.enable : No 1
4063 provides the same functionality as above for the IPv6 case.
4064 .It Va net.inet.ip.fw.one_pass : No 1
4065 When set, the packet exiting from the
4069 node is not passed though the firewall again.
4070 Otherwise, after an action, the packet is
4071 reinjected into the firewall at the next rule.
4072 .It Va net.inet.ip.fw.tables_max : No 128
4073 Maximum number of tables.
4074 .It Va net.inet.ip.fw.verbose : No 1
4075 Enables verbose messages.
4076 .It Va net.inet.ip.fw.verbose_limit : No 0
4077 Limits the number of messages produced by a verbose firewall.
4078 .It Va net.inet6.ip6.fw.deny_unknown_exthdrs : No 1
4079 If enabled packets with unknown IPv6 Extension Headers will be denied.
4080 .It Va net.link.ether.ipfw : No 0
4081 Controls whether layer-2 packets are passed to
4084 .It Va net.link.bridge.ipfw : No 0
4085 Controls whether bridged packets are passed to
4088 .It Va net.inet.ip.fw.nat64_debug : No 0
4089 Controls debugging messages produced by
4092 .It Va net.inet.ip.fw.nat64_direct_output : No 0
4093 Controls the output method used by
4096 .Bl -tag -width indent
4098 A packet is handled by
4101 First time an original packet is handled by
4106 Then translated packet is queued via netisr to input processing again.
4108 A packet is handled by
4110 only once, and after translation it will be pushed directly to outgoing
4114 .Sh INTERNAL DIAGNOSTICS
4115 There are some commands that may be useful to understand current state
4116 of certain subsystems inside kernel module.
4117 These commands provide debugging output which may change without notice.
4119 Currently the following commands are available as
4122 .Bl -tag -width indent
4124 Lists all interface which are currently tracked by
4126 with their in-kernel status.
4128 List all table lookup algorithms currently available.
4131 There are far too many possible uses of
4133 so this Section will only give a small set of examples.
4135 .Ss BASIC PACKET FILTERING
4136 This command adds an entry which denies all tcp packets from
4137 .Em cracker.evil.org
4138 to the telnet port of
4140 from being forwarded by the host:
4142 .Dl "ipfw add deny tcp from cracker.evil.org to wolf.tambov.su telnet"
4144 This one disallows any connection from the entire cracker's
4147 .Dl "ipfw add deny ip from 123.45.67.0/24 to my.host.org"
4149 A first and efficient way to limit access (not using dynamic rules)
4150 is the use of the following rules:
4152 .Dl "ipfw add allow tcp from any to any established"
4153 .Dl "ipfw add allow tcp from net1 portlist1 to net2 portlist2 setup"
4154 .Dl "ipfw add allow tcp from net3 portlist3 to net3 portlist3 setup"
4156 .Dl "ipfw add deny tcp from any to any"
4158 The first rule will be a quick match for normal TCP packets,
4159 but it will not match the initial SYN packet, which will be
4162 rules only for selected source/destination pairs.
4163 All other SYN packets will be rejected by the final
4167 If you administer one or more subnets, you can take advantage
4168 of the address sets and or-blocks and write extremely
4169 compact rulesets which selectively enable services to blocks
4170 of clients, as below:
4172 .Dl "goodguys=\*q{ 10.1.2.0/24{20,35,66,18} or 10.2.3.0/28{6,3,11} }\*q"
4173 .Dl "badguys=\*q10.1.2.0/24{8,38,60}\*q"
4175 .Dl "ipfw add allow ip from ${goodguys} to any"
4176 .Dl "ipfw add deny ip from ${badguys} to any"
4177 .Dl "... normal policies ..."
4181 option could be used to do automated anti-spoofing by adding the
4182 following to the top of a ruleset:
4184 .Dl "ipfw add deny ip from any to any not verrevpath in"
4186 This rule drops all incoming packets that appear to be coming to the
4187 system on the wrong interface.
4188 For example, a packet with a source
4189 address belonging to a host on a protected internal network would be
4190 dropped if it tried to enter the system from an external interface.
4194 option could be used to do similar but more restricted anti-spoofing
4195 by adding the following to the top of a ruleset:
4197 .Dl "ipfw add deny ip from any to any not antispoof in"
4199 This rule drops all incoming packets that appear to be coming from another
4200 directly connected system but on the wrong interface.
4201 For example, a packet with a source address of
4202 .Li 192.168.0.0/24 ,
4211 option could be used to (re)mark user traffic,
4212 by adding the following to the appropriate place in ruleset:
4214 .Dl "ipfw add setdscp be ip from any to any dscp af11,af21"
4215 .Ss SELECTIVE MIRRORING
4216 If your network has network traffic analyzer
4217 connected to your host directly via dedicated interface
4218 or remotely via RSPAN vlan, you can selectively mirror
4219 some Ethernet layer2 frames to the analyzer.
4221 First, make sure your firewall is already configured and runs.
4222 Then, enable layer2 processing if not already enabled:
4224 .Dl "sysctl net.link.ether.ipfw=1"
4226 Next, load needed additional kernel modules:
4228 .Dl "kldload ng_ether ng_ipfw"
4230 Optionally, make system load these modules automatically
4233 .Dl sysrc kld_list+="ng_ether ng_ipfw"
4237 kernel module to transmit mirrored copies of layer2 frames
4238 out via vlan900 interface:
4240 .Dl "ngctl connect ipfw: vlan900: 1 lower"
4242 Think of "1" here as of "mirroring instance index" and vlan900 is its
4244 You can have arbitrary number of instances.
4249 At last, actually start mirroring of selected frames using "instance 1".
4250 For frames incoming from em0 interface:
4252 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 in recv em0"
4254 For frames outgoing to em0 interface:
4256 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 out xmit em0"
4258 For both incoming and outgoing frames while flowing through em0:
4260 .Dl "ipfw add ngtee 1 ip from any to 192.168.0.1 layer2 via em0"
4262 Make sure you do not perform mirroring for already duplicated frames
4263 or kernel may hang as there is no safety net.
4265 In order to protect a site from flood attacks involving fake
4266 TCP packets, it is safer to use dynamic rules:
4268 .Dl "ipfw add check-state"
4269 .Dl "ipfw add deny tcp from any to any established"
4270 .Dl "ipfw add allow tcp from my-net to any setup keep-state"
4272 This will let the firewall install dynamic rules only for
4273 those connection which start with a regular SYN packet coming
4274 from the inside of our network.
4275 Dynamic rules are checked when encountering the first
4284 rule should usually be placed near the beginning of the
4285 ruleset to minimize the amount of work scanning the ruleset.
4286 Your mileage may vary.
4288 For more complex scenarios with dynamic rules
4292 can be used to precisely control creation and checking of dynamic rules.
4293 Example of usage of these options are provided in
4294 .Sx NETWORK ADDRESS TRANSLATION (NAT)
4297 To limit the number of connections a user can open
4298 you can use the following type of rules:
4300 .Dl "ipfw add allow tcp from my-net/24 to any setup limit src-addr 10"
4301 .Dl "ipfw add allow tcp from any to me setup limit src-addr 4"
4303 The former (assuming it runs on a gateway) will allow each host
4304 on a /24 network to open at most 10 TCP connections.
4305 The latter can be placed on a server to make sure that a single
4306 client does not use more than 4 simultaneous connections.
4309 stateful rules can be subject to denial-of-service attacks
4310 by a SYN-flood which opens a huge number of dynamic rules.
4311 The effects of such attacks can be partially limited by
4314 variables which control the operation of the firewall.
4316 Here is a good usage of the
4318 command to see accounting records and timestamp information:
4322 or in short form without timestamps:
4326 which is equivalent to:
4330 Next rule diverts all incoming packets from 192.168.2.0/24
4331 to divert port 5000:
4333 .Dl ipfw divert 5000 ip from 192.168.2.0/24 to any in
4335 The following rules show some of the applications of
4339 for simulations and the like.
4341 This rule drops random incoming packets with a probability
4344 .Dl "ipfw add prob 0.05 deny ip from any to any in"
4346 A similar effect can be achieved making use of
4350 .Dl "ipfw add pipe 10 ip from any to any"
4351 .Dl "ipfw pipe 10 config plr 0.05"
4353 We can use pipes to artificially limit bandwidth, e.g.\& on a
4354 machine acting as a router, if we want to limit traffic from
4355 local clients on 192.168.2.0/24 we do:
4357 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4358 .Dl "ipfw pipe 1 config bw 300Kbit/s queue 50KBytes"
4360 note that we use the
4362 modifier so that the rule is not used twice.
4363 Remember in fact that
4365 rules are checked both on incoming and outgoing packets.
4367 Should we want to simulate a bidirectional link with bandwidth
4368 limitations, the correct way is the following:
4370 .Dl "ipfw add pipe 1 ip from any to any out"
4371 .Dl "ipfw add pipe 2 ip from any to any in"
4372 .Dl "ipfw pipe 1 config bw 64Kbit/s queue 10Kbytes"
4373 .Dl "ipfw pipe 2 config bw 64Kbit/s queue 10Kbytes"
4375 The above can be very useful, e.g.\& if you want to see how
4376 your fancy Web page will look for a residential user who
4377 is connected only through a slow link.
4378 You should not use only one pipe for both directions, unless
4379 you want to simulate a half-duplex medium (e.g.\& AppleTalk,
4381 It is not necessary that both pipes have the same configuration,
4382 so we can also simulate asymmetric links.
4384 Should we want to verify network performance with the RED queue
4385 management algorithm:
4387 .Dl "ipfw add pipe 1 ip from any to any"
4388 .Dl "ipfw pipe 1 config bw 500Kbit/s queue 100 red 0.002/30/80/0.1"
4390 Another typical application of the traffic shaper is to
4391 introduce some delay in the communication.
4392 This can significantly affect applications which do a lot of Remote
4393 Procedure Calls, and where the round-trip-time of the
4394 connection often becomes a limiting factor much more than
4397 .Dl "ipfw add pipe 1 ip from any to any out"
4398 .Dl "ipfw add pipe 2 ip from any to any in"
4399 .Dl "ipfw pipe 1 config delay 250ms bw 1Mbit/s"
4400 .Dl "ipfw pipe 2 config delay 250ms bw 1Mbit/s"
4402 Per-flow queueing can be useful for a variety of purposes.
4403 A very simple one is counting traffic:
4405 .Dl "ipfw add pipe 1 tcp from any to any"
4406 .Dl "ipfw add pipe 1 udp from any to any"
4407 .Dl "ipfw add pipe 1 ip from any to any"
4408 .Dl "ipfw pipe 1 config mask all"
4410 The above set of rules will create queues (and collect
4411 statistics) for all traffic.
4412 Because the pipes have no limitations, the only effect is
4413 collecting statistics.
4414 Note that we need 3 rules, not just the last one, because
4417 tries to match IP packets it will not consider ports, so we
4418 would not see connections on separate ports as different
4421 A more sophisticated example is limiting the outbound traffic
4422 on a net with per-host limits, rather than per-network limits:
4424 .Dl "ipfw add pipe 1 ip from 192.168.2.0/24 to any out"
4425 .Dl "ipfw add pipe 2 ip from any to 192.168.2.0/24 in"
4426 .Dl "ipfw pipe 1 config mask src-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4427 .Dl "ipfw pipe 2 config mask dst-ip 0x000000ff bw 200Kbit/s queue 20Kbytes"
4429 In the following example, we need to create several traffic bandwidth
4430 classes and we need different hosts/networks to fall into different classes.
4431 We create one pipe for each class and configure them accordingly.
4432 Then we create a single table and fill it with IP subnets and addresses.
4433 For each subnet/host we set the argument equal to the number of the pipe
4435 Then we classify traffic using a single rule:
4437 .Dl "ipfw pipe 1 config bw 1000Kbyte/s"
4438 .Dl "ipfw pipe 4 config bw 4000Kbyte/s"
4440 .Dl "ipfw table T1 create type addr"
4441 .Dl "ipfw table T1 add 192.168.2.0/24 1"
4442 .Dl "ipfw table T1 add 192.168.0.0/27 4"
4443 .Dl "ipfw table T1 add 192.168.0.2 1"
4445 .Dl "ipfw add pipe tablearg ip from 'table(T1)' to any"
4449 action, the table entries may include hostnames and IP addresses.
4451 .Dl "ipfw table T2 create type addr ftype ip"
4452 .Dl "ipfw table T2 add 192.168.2.0/24 10.23.2.1"
4453 .Dl "ipfw table T21 add 192.168.0.0/27 router1.dmz"
4455 .Dl "ipfw add 100 fwd tablearg ip from any to table(1)"
4457 In the following example per-interface firewall is created:
4459 .Dl "ipfw table IN create type iface valtype skipto,fib"
4460 .Dl "ipfw table IN add vlan20 12000,12"
4461 .Dl "ipfw table IN add vlan30 13000,13"
4462 .Dl "ipfw table OUT create type iface valtype skipto"
4463 .Dl "ipfw table OUT add vlan20 22000"
4464 .Dl "ipfw table OUT add vlan30 23000"
4466 .Dl "ipfw add 100 setfib tablearg ip from any to any recv 'table(IN)' in"
4467 .Dl "ipfw add 200 skipto tablearg ip from any to any recv 'table(IN)' in"
4468 .Dl "ipfw add 300 skipto tablearg ip from any to any xmit 'table(OUT)' out"
4470 The following example illustrate usage of flow tables:
4472 .Dl "ipfw table fl create type flow:src-ip,proto,dst-ip,dst-port"
4473 .Dl "ipfw table fl add 2a02:6b8:77::88,tcp,2a02:6b8:77::99,80 11"
4474 .Dl "ipfw table fl add 10.0.0.1,udp,10.0.0.2,53 12"
4476 .Dl "ipfw add 100 allow ip from any to any flow 'table(fl,11)' recv ix0"
4478 To add a set of rules atomically, e.g.\& set 18:
4480 .Dl "ipfw set disable 18"
4481 .Dl "ipfw add NN set 18 ... # repeat as needed"
4482 .Dl "ipfw set enable 18"
4484 To delete a set of rules atomically the command is simply:
4486 .Dl "ipfw delete set 18"
4488 To test a ruleset and disable it and regain control if something goes wrong:
4490 .Dl "ipfw set disable 18"
4491 .Dl "ipfw add NN set 18 ... # repeat as needed"
4492 .Dl "ipfw set enable 18; echo done; sleep 30 && ipfw set disable 18"
4494 Here if everything goes well, you press control-C before the "sleep"
4495 terminates, and your ruleset will be left active.
4496 Otherwise, e.g.\& if
4497 you cannot access your box, the ruleset will be disabled after
4498 the sleep terminates thus restoring the previous situation.
4500 To show rules of the specific set:
4502 .Dl "ipfw set 18 show"
4504 To show rules of the disabled set:
4506 .Dl "ipfw -S set 18 show"
4508 To clear a specific rule counters of the specific set:
4510 .Dl "ipfw set 18 zero NN"
4512 To delete a specific rule of the specific set:
4514 .Dl "ipfw set 18 delete NN"
4515 .Ss NAT, REDIRECT AND LSNAT
4516 First redirect all the traffic to nat instance 123:
4518 .Dl "ipfw add nat 123 all from any to any"
4520 Then to configure nat instance 123 to alias all the outgoing traffic with ip
4521 192.168.0.123, blocking all incoming connections, trying to keep
4522 same ports on both sides, clearing aliasing table on address change
4523 and keeping a log of traffic/link statistics:
4525 .Dl "ipfw nat 123 config ip 192.168.0.123 log deny_in reset same_ports"
4527 Or to change address of instance 123, aliasing table will be cleared (see
4530 .Dl "ipfw nat 123 config ip 10.0.0.1"
4532 To see configuration of nat instance 123:
4534 .Dl "ipfw nat 123 show config"
4536 To show logs of all the instances in range 111-999:
4538 .Dl "ipfw nat 111-999 show"
4540 To see configurations of all instances:
4542 .Dl "ipfw nat show config"
4544 Or a redirect rule with mixed modes could looks like:
4546 .Dl "ipfw nat 123 config redirect_addr 10.0.0.1 10.0.0.66"
4547 .Dl " redirect_port tcp 192.168.0.1:80 500"
4548 .Dl " redirect_proto udp 192.168.1.43 192.168.1.1"
4549 .Dl " redirect_addr 192.168.0.10,192.168.0.11"
4550 .Dl " 10.0.0.100 # LSNAT"
4551 .Dl " redirect_port tcp 192.168.0.1:80,192.168.0.10:22"
4554 or it could be split in:
4556 .Dl "ipfw nat 1 config redirect_addr 10.0.0.1 10.0.0.66"
4557 .Dl "ipfw nat 2 config redirect_port tcp 192.168.0.1:80 500"
4558 .Dl "ipfw nat 3 config redirect_proto udp 192.168.1.43 192.168.1.1"
4559 .Dl "ipfw nat 4 config redirect_addr 192.168.0.10,192.168.0.11,192.168.0.12"
4561 .Dl "ipfw nat 5 config redirect_port tcp"
4562 .Dl " 192.168.0.1:80,192.168.0.10:22,192.168.0.20:25 500"
4564 Sometimes you may want to mix NAT and dynamic rules.
4565 It could be achieved with
4570 Problem is, you need to create dynamic rule before NAT and check it
4571 after NAT actions (or vice versa) to have consistent addresses and ports.
4574 option will trigger activation of existing dynamic state, and action of such
4575 rule will be performed as soon as rule is matched.
4578 rule packet need to be passed to NAT, not allowed as soon is possible.
4580 There is example of set of rules to achieve this.
4581 Bear in mind that this is example only and it is not very useful by itself.
4583 On way out, after all checks place this rules:
4585 .Dl "ipfw add allow record-state skip-action"
4586 .Dl "ipfw add nat 1"
4588 And on way in there should be something like this:
4590 .Dl "ipfw add nat 1"
4591 .Dl "ipfw add check-state"
4593 Please note, that first rule on way out doesn't allow packet and doesn't
4594 execute existing dynamic rules.
4595 All it does, create new dynamic rule with
4597 action, if it is not created yet.
4598 Later, this dynamic rule is used on way in by
4601 .Ss CONFIGURING CODEL, PIE, FQ-CODEL and FQ-PIE AQM
4605 AQM can be configured for
4615 AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4618 .Dl "ipfw pipe 1 config bw 1mbits/s codel"
4619 .Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4625 AQM using different configurations parameters for traffic from
4626 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4628 .Dl "ipfw pipe 1 config bw 1mbits/s"
4629 .Dl "ipfw queue 1 config pipe 1 codel target 8ms interval 160ms ecn"
4630 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4636 AQM using default configuration for traffic from 192.168.0.0/24 and 1Mbits/s
4639 .Dl "ipfw pipe 1 config bw 1mbits/s pie"
4640 .Dl "ipfw add 100 pipe 1 ip from 192.168.0.0/24 to any"
4646 AQM using different configuration parameters for traffic from
4647 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4649 .Dl "ipfw pipe 1 config bw 1mbits/s"
4650 .Dl "ipfw queue 1 config pipe 1 pie target 20ms tupdate 30ms ecn"
4651 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4656 AQM can be configured for
4662 scheduler using different configurations parameters for traffic from
4663 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4665 .Dl "ipfw pipe 1 config bw 1mbits/s"
4666 .Dl "ipfw sched 1 config pipe 1 type fq_codel"
4667 .Dl "ipfw queue 1 config sched 1"
4668 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4672 default configuration for a
4674 such as disable ECN and change the
4678 .Dl "ipfw sched 1 config pipe 1 type fq_codel target 10ms noecn"
4684 scheduler using different configurations parameters for traffic from
4685 192.168.0.0/24 and 1Mbits/s rate limit, we do:
4687 .Dl "ipfw pipe 1 config bw 1mbits/s"
4688 .Dl "ipfw sched 1 config pipe 1 type fq_pie"
4689 .Dl "ipfw queue 1 config sched 1"
4690 .Dl "ipfw add 100 queue 1 ip from 192.168.0.0/24 to any"
4692 The configurations of
4695 can be changed in a similar way as for
4719 utility first appeared in
4724 Stateful extensions were introduced in
4727 was introduced in Summer 2002.
4729 .An Ugen J. S. Antsilevich ,
4730 .An Poul-Henning Kamp ,
4734 .An Rasool Al-Saadi .
4737 API based upon code written by
4741 Dummynet has been introduced by Luigi Rizzo in 1997-1998.
4743 Some early work (1999-2000) on the
4745 traffic shaper supported by Akamba Corp.
4747 The ipfw core (ipfw2) has been completely redesigned and
4748 reimplemented by Luigi Rizzo in summer 2002.
4751 options have been added by various developers over the years.
4754 In-kernel NAT support written by
4755 .An Paolo Pisati Aq Mt piso@FreeBSD.org
4756 as part of a Summer of Code 2005 project.
4760 support has been developed by
4761 .An The Centre for Advanced Internet Architectures (CAIA) Aq http://www.caia.swin.edu.au .
4762 The primary developers and maintainers are David Hayes and Jason But.
4763 For further information visit:
4764 .Aq http://www.caia.swin.edu.au/urp/SONATA
4766 Delay profiles have been developed by Alessandro Cerri and
4767 Luigi Rizzo, supported by the
4768 European Commission within Projects Onelab and Onelab2.
4770 CoDel, PIE, FQ-CoDel and FQ-PIE AQM for Dummynet have been implemented by
4771 .An The Centre for Advanced Internet Architectures (CAIA)
4772 in 2016, supported by The Comcast Innovation Fund.
4773 The primary developer is
4776 The syntax has grown over the years and sometimes it might be confusing.
4777 Unfortunately, backward compatibility prevents cleaning up mistakes
4778 made in the definition of the syntax.
4782 Misconfiguring the firewall can put your computer in an unusable state,
4783 possibly shutting down network services and requiring console access to
4784 regain control of it.
4786 Incoming packet fragments diverted by
4788 are reassembled before delivery to the socket.
4789 The action used on those packet is the one from the
4790 rule which matches the first fragment of the packet.
4792 Packets diverted to userland, and then reinserted by a userland process
4793 may lose various packet attributes.
4794 The packet source interface name
4795 will be preserved if it is shorter than 8 bytes and the userland process
4796 saves and reuses the sockaddr_in
4799 otherwise, it may be lost.
4800 If a packet is reinserted in this manner, later rules may be incorrectly
4801 applied, making the order of
4803 rules in the rule sequence very important.
4805 Dummynet drops all packets with IPv6 link-local addresses.
4811 may not behave as expected.
4812 In particular, incoming SYN packets may
4813 have no uid or gid associated with them since they do not yet belong
4814 to a TCP connection, and the uid/gid associated with a packet may not
4815 be as expected if the associated process calls
4817 or similar system calls.
4819 Rule syntax is subject to the command line environment and some patterns
4820 may need to be escaped with the backslash character
4821 or quoted appropriately.
4823 Due to the architecture of
4825 ipfw nat is not compatible with the TCP segmentation offloading (TSO).
4826 Thus, to reliably nat your network traffic, please disable TSO
4830 ICMP error messages are not implicitly matched by dynamic rules
4831 for the respective conversations.
4832 To avoid failures of network error detection and path MTU discovery,
4833 ICMP error messages may need to be allowed explicitly through static
4840 actions may lead to confusing behaviour if ruleset has mistakes,
4841 and/or interaction with other subsystems (netgraph, dummynet, etc.) is used.
4842 One possible case for this is packet leaving
4844 in subroutine on the input pass, while later on output encountering unpaired
4847 As the call stack is kept intact after input pass, packet will suddenly
4848 return to the rule number used on input pass, not on output one.
4849 Order of processing should be checked carefully to avoid such mistakes.